QUANTUM COMPUTING TALENT MANAGEMENT EDUCATING IN THE RESHAPING BASEBALL ENHANCING SUPPLY

And its overlap Enabling top talent – AGE OF AI WITH ANALYTICS & TECH CHAIN RESILIENCE
with operations not replacing it – with Future-proofing Baseball in its data-driven Mitigating hurricane-
research (O.R.) smart computers O.R. practitioners decision-making era driven disruptions

2025 I Volume 52, Issue 2

OR/MS TODAY ADVERTISING EDITORS OF OTHER
AND EDITORIAL OFFICE INFORMS PUBLICATIONS
5521 Research Park Drive, Suite 200, Catonsville, MD 21228 Decision Analysis
Telephone: 443.757.3500 Vicki M. Bier, University of Wisconsin-Madison
Email: [email protected] Information Systems Research
Editor Suprateek Sarker, University of Virginia
Kara Tucker, [email protected] INFORMS Analytics Collections
Director of Marketing Nicholas G. Hall, The Ohio State University
Mary Leszczynski, [email protected] INFORMS Journal on Applied Analytics
Director of Publications Alexandra Newman, Colorado School of Mines
Matthew Walls, [email protected] INFORMS Journal on Computing
Advertising Sales Alice E. Smith, Auburn University
[email protected] INFORMS Journal on Data Science
Production Yu Ding, Georgia Institute of Technology
Kristine Ferg, [email protected] INFORMS Journal on Optimization
Magazine Editorial Advisory Board Oktay Günlük, Georgia Institute of Technology
Anna Nagurney, Committee Chair INFORMS Transactions on Education
INFORMS Online Stefan Creemers, IÉSEG
www.informs.org Management Science
ormstoday.informs.org Christoph Loch, Cambridge University
Manufacturing & Service Operations Management
INFORMS BOARD OF DIRECTORS Georgia Perakis, Massachusetts Institute of Technology
President Marketing Science
Dave Hunt, Oliver Wyman Puneet Manchanda, University of Michigan
President-Elect Mathematics of Operations Research
Mark Lewis, Cornell University Katya Scheinberg, Georgia Institute of Technology
Past President Operations Research
Julie Swann, North Carolina State University Amy Ward, University of Chicago
Secretary Organization Science
Pelin Pekgun, Wake Forest University Lamar Pierce, Washington University in St. Louis
Treasurer Service Science
Susan Martonosi, Harvey Mudd College Guillaume Roels, INSEAD
Vice President, Chapters and Forums Stochastic Systems
Vinod Cheriyan, Stitch Fix Devavrat Shah, Massachusetts Institute of Technology
Vice President, Education Strategy Science
Illya Hicks, Rice University Todd Zenger, University of Utah
Vice President, International Activities Transportation Science
Miguel Anjos, University of Edinburgh Karen Smilowitz, Northwestern University
Vice President, Marketing, Communications TutORials in Operations Research
and Outreach David L. Alderson, Naval Postgraduate School
Tinglong Dai, Johns Hopkins University
Vice President, Meetings
Ariela Sofer, George Mason University OR/MS Today (ISSN 1085-1038) is published quarterly by the Institute for
Operations Research and the Management Sciences (INFORMS). Canada
Vice President, Membership and Post International Publications Mail (Canadian Distribution) Sales Agreement
Professional Recognition No. 1220047. Deadlines for contributions: Manuscripts and news items
Anahita Khojandi, University of Tennessee should arrive no later than six weeks prior to the first day of the month
of publication. Address correspondence regarding editorial content to
Vice President, Practice
the editor, Kara Tucker: email: [email protected]; phone: 443-757-3572.
Robin Lougee, KnitWell
The opinions expressed in OR/MS Today are those of the authors and do not
Vice President, Publications necessarily reflect the opinions of INFORMS, its officers or the editorial staff
Christopher Tang, University of California, Los Angeles of OR/MS Today. Membership subscriptions for OR/MS Today are included
in annual dues. INFORMS offers nonmember subscriptions to institutions;
Vice President, Sections and Societies the rate is $62 for USA, $85 all other countries. Single copies can be
Theresa Roeder, San Francisco University purchased for $25 plus postage. Periodicals postage paid at Catonsville, MD,
Vice President, Technology Strategy and additional mailing offices. Printed in the United States of America.
Warren Hearnes, OptiML AI POSTMASTER: Send address changes to OR/MS Today, INFORMS, 5521
Research Park Dr., Suite 200, Catonsville, MD 21228. OR/MS Today ©2025
Executive Director by the Institute for Operations Research and the Management Sciences.
Elena Gerstmann, Ph.D. All rights reserved.
 Introducing

Gurobi 12.0
Every Solution, Globally Optimized

With Gurobi 12.0, users can solve complex nonlinear problems to global
optimality, and achieve superior performance across all mixed-integer
optimizations thanks in part to more efficient use of CPU cores.

• Unmatched performance • Smarter resource management

• Innovative nonlinear capabilities • Revamped documentation

Learn more about Gurobi 12.0 or try it for free at

gurobi.com/12.0
2025 ISSUE 2 CONTENTS

20 24 30 36
QUANTUM COMPUTING TALENT MANAGEMENT EDUCATING IN THE RESHAPING BASEBALL
What can quantum ‘Here come the AGE OF AI WITH ANALYTICS & TECH
computers do for you? digital workers!’ Future-proofing A scientific era
And what can you do Enabling top talent O.R. practitioners. defined by data-driven
for them? with smart computers, decision-making.
BY MICHAEL FRY,
BY DAVID E. BERNAL NEIRA, instead of replacing them. HARISH KRISHNAN, BY ZIYU (JACK) CHEN,
REBEKAH HERRMAN, BY HALUK DEMIRKAN MARIA VLASIOU, TRISTAN ROBBINS,
MOHAMMADHOSSEIN AND JIM SPOHRER BRADLEY WARNER WARNER GEPHARDT AND
MOHAMMADISIAHROUDI, WENJUN ZHOU, SCOTT NESTLER, CAP-X
GIACOMO NANNICINI, ET AL. ANNE G. ROBINSON AND
RANGANATH NUGGEHALLI

42 DEPARTMENTS
ENHANCING SUPPLY 4 Inside Story 16 Viewpoint 54 Member Insights
CHAIN RESILIENCE
6 President's Desk 18 Letter to Self 56 Conference Preview
Strategies for
mitigating hurricane- 8 HQ Highlights 46 Roundtable Profile 58 INFORMS Awards
driven disruptions and Prizes
in breakbulk logistics. 11 INFORMS Online 50 INFORMS Initiatives
60 Student Perspectives
BY DENNIS HERNANDEZ 12 Advocacy in D.C. 52 President-Elect
AND GEMMA BERENGUER Candidate 63 Last Word
14 Forum Statements

2 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 Solve your Mathematical Optimization
problems in minutes

Hexaly Optimizer is the world’s fastest optimization

solver for Routing, Scheduling, Packing, and more.

200 COMPANIES TRUST US

Join a fast-growing community of 1,000 developers.
They leveraged Hexaly to build 400 apps for 20,000 users. Why not you?

Try it for free

hexaly.com/signup

hexaly.com
INSIDE STORY

FUTURE-READY:
EMBRACING
AI-DRIVEN
TRANSFORMATIONS
BY KARA TUCKER

IN EVERY ERA, PROFESSIONALS MUST the Digital Workers!’” (p. 24), automation and
adapt to the changing tides of innovation – but few intelligent agents are becoming integral members
periods have seen such sweeping transformation of modern organizations. From robotic process
as the one we’re navigating now. From artificial automation in back offices to virtual assistants in
intelligence (AI) and quantum computing to frontline operations, digital workers are not just
advanced analytics and automation, the 21st supporting human teams – they are redefining
In this issue, century is redefining not only how we work but roles, workflows and management structures. The
we explore also how we think, learn and compete. In this challenge facing today’s leaders is not whether
issue, we explore how operations research (O.R.) to integrate these technologies, but how to
how operations and analytics professionals across disciplines orchestrate human-digital collaboration in ways
research (O.R.) are preparing themselves for a future that’s that enhance both efficiency and creativity.
already here. Alongside opportunity, however, disruption
and analytics At the heart of this transformation is the looms large – especially in sectors like logistics,
professionals need to educate for resilience and relevance. where external shocks can quickly ripple through
In “Educating in the Age of AI: Future-proofing global supply chains. On page 42, in “Enhancing
across O.R. Practitioners” (p. 30), the authors discuss Supply Chain Resilience: Strategies to Mitigate
how academic institutions and professional Risk of Hurricane-driven Disruptions in Breakbulk
disciplines development programs are retooling their Logistics,” the authors dive into how professionals
are preparing curricula to emphasize not just technical are using predictive analytics, scenario modeling
expertise, but also adaptability, interdisciplinary and real-time data to better prepare for climate-
themselves for thinking and ethical awareness. The O.R. driven uncertainty. What emerges is a clear message:
a future that’s professional of tomorrow must be as fluent agility and foresight are no longer optional – they
in AI-driven decision tools as they are in are central to competitive survival.
already here. human-centered design and systems thinking. Although many of today’s technologies
Future-proofing isn’t just about keeping pace – are incremental, others are revolutionary in
it’s about cultivating a mindset that’s ready scope. Take quantum computing. On page 20,
for continuous evolution. members of the INFORMS Quantum Computing
Nowhere is this evolution more apparent and Operations Research Ad Hoc Committee
than in the emerging workforce landscape. As demystify a technology that has long seemed like
highlighted in “Talent Management: ‘Here Come science fiction but is edging closer to practical

4 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

reality. Quantum algorithms promise exponential organization, navigating global risk, testing the
speed-ups for specific problems in optimization, limits of quantum theory or optimizing a baseball
simulation and chemical engineering – offering lineup, success hinges on the ability to marry deep
tantalizing possibilities for O.R. professionals analytical rigor with ethical foresight and human-
in sectors as diverse as energy, logistics and centric design. As practitioners,
pharmaceuticals. The field may be nascent,
but its potential impact is profound.
INFORMS members sit at the center of this
transformation. As practitioners, educators,
educators,
And yet, the technological transformation is not researchers and leaders, you’re shaping how these researchers
confined to traditional industries. Even the world technologies are implemented, interpreted and
of sports is feeling the tremors of analytics-driven improved. Your work ensures that we don’t just
and leaders,
change. Baseball is currently in its data-driven chase innovation for its own sake, but harness it INFORMS
decision-making era. “Reshaping Baseball: The to solve real problems, drive equitable growth and
Impact of Analytics and Technology” (p. 36) build a more resilient world. members
offers a compelling look at how analytics and This issue is a celebration of that work – and are shaping
data science are reshaping how teams scout a call to action. I encourage you to explore each
talent, optimize pitching mechanics, strategize article not only for insights, but for the questions how these
during games and engage with fans. There are they raise. How can you apply these lessons in your technologies
even AI umps! The democratization of data, own practice? What skills will your teams need
combined with advanced modeling techniques, next year that they don’t have today? How can we are improved,
is leveling the playing field and shifting the art collectively ensure that these powerful tools are
of the game toward science – without losing the used responsibly, inclusively and effectively?
interpreted and
soul of competition. The future is no longer something we plan implemented.
What binds all these threads together is a for – it’s something this community is already
shared imperative: the need to lead with insight, shaping, together.
responsibility and vision. Whether you’re deploying
AI in the classroom, rethinking roles in your KARA TUCKER is editor of OR/MS Today and Analytics magazines.

DONATE TO THE FUND OF YOUR CHOICE

General Fund: Empowers INFORMS Student Fund: Helps students gain
to respond to emerging needs, the experience and exposure that
launch new initiatives, and support comes from attending and presenting
innovative programs that strengthen at INFORMS conferences by defraying
the operations research and travel costs.
analytics profession.
https://www.informs.org/Donate
PRESIDENT'S DESK

THE POWER OF
COLLABORATION
BY DAVID HUNT

PROFESSIONAL ASSOCIATIONS PROVIDE countries and cultures; among disciplines; through

invaluable services to their members and to society generations; between academia, industry and
at large. These services include promoting and government; and with other associations.
advancing community interests, disseminating
current research, creating educational and Collaboration Across Countries and Cultures
certification opportunities, offering career INFORMS is based in the United States, but it is
services, and celebrating member achievements. important to recognize that more than 30% of
Associations also play a vital role in advocating our members reside outside the U.S., representing
for their fields, upholding ethical and scientific a vast community spread across more than 80
standards, and fostering inclusive, global countries. The distribution agreement INFORMS
communities. However, perhaps none of these signed with EBSCO extends our journals to over
services is more critical than fostering collaboration 130 countries, further enhancing our global reach.
among members and across various sectors. We are committed to strengthening our ties with
Having recently returned from a successful global members and building a robust worldwide
INFORMS Analytics+ Conference, I am struck community because we understand that working
by the multitude of ideas generated and the together helps the world make smarter decisions.
incredible advancements that emerge through This commitment to international collaboration
the power of collaboration. The event served not only enriches our organization but also
as a testament to the innovative potential enhances the quality of research and practice in
unlocked when professionals come together to operations research (O.R.) and analytics. INFORMS
share insights and expertise. I am also acutely has enjoyed a strong global presence at our U.S.-
aware of the important role that INFORMS based meetings and is committed to hosting more
plays in facilitating collaboration and the meetings in other countries, including the meeting
tremendous opportunity – and indeed, the last year in Medellín, Colombia, and the upcoming
necessity – for INFORMS to enhance its efforts meeting in Singapore.
to support collaboration. This need is especially One specific example was raised by 2025
pronounced as the United States, along with Analytics+ keynote speaker Dick den Hertog
many other countries, increasingly prioritizes from the University of Amsterdam. Professor
domestic interests over global cooperation. Hertog estimates that there are 3.5 million data
At the same time, collaboration has taken on professionals focused on social impact projects
heightened significance for our profession in in low- and middle-income countries (LMICs),
light of ongoing challenges, including mounting many of whom work in isolation. He asked if
threats to academic freedom, punitive actions INFORMS could play a role in connecting this
and restrictions targeting international group to support better collaboration efforts. This
students and employees, and uncertainty need is especially pressing, as some government
in research funding, among broader funding for these initiatives has been cut. We are
economic concerns. actively exploring how INFORMS can contribute
There are numerous opportunities for INFORMS to this cause and support collaboration among this
to further support collaboration in its various growing O.R. and analytics community. Please
forms, including fostering collaboration across reach out to me if you would like to be involved.

6 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

Collaboration Across Disciplines Currently, there is an ad hoc committee looking
Operations research at its core, to paraphrase two at expanding regional meetings, and the INFORMS
of its founders, is a “scientific method of making Chapters and Forums Committee is exploring
decisions regarding operations” [1]. Achieving this how to better support local chapters. Creating
requires embracing a wide array of disciplines. more opportunities for academic and industry
As outlined in our Strategic Plan, INFORMS members to meet will spur more collaboration
encompasses and values the full breadth of and innovation.
disciplines represented in our community [2].
These include operations research, management Collaboration Among Associations
science, analytics, artificial intelligence, data INFORMS also has opportunities to collaborate
science, industrial and systems engineering, with other STEM-based associations to advocate
statistics, applied mathematics, economics, for topics of mutual interest. Earlier this year,
behavioral science, quantum computing and INFORMS signed a joint letter calling for the
many other related fields. reinstatement of government databases that
INFORMS is uniquely positioned to connect had been removed. We are also working with the
the best ideas from this broad range of disciplines American Statistical Association and others to
to drive innovation and solve complex problems promote the Data Science and Literacy Act [3].
more effectively. We will continue to nurture Discussions are ongoing to identify potential
these relationships and build cross-disciplinary advocacy efforts and actions of common interest
collaborations, as we have done through that will benefit both INFORMS members and the As a global
recent activities jointly undertaken with the broader scientific community.
Association for the Advancement of Artificial community,
Intelligence (AAAI). O.R. Founded on Collaboration
It is no coincidence that the field of operations
INFORMS
Collaboration Across and research was born out of international fosters
Among Generations collaboration. During World War II, the
Mentoring is consistently highlighted by early- Nobel Prize-winning physicist Patrick collaboration
career members as beneficial for their professional Blackett was appointed director of British across
growth. INFORMS currently offers a Mentor Match Naval Operational Research and, along with
program pairing an early-career member with a his colleagues in England, began pioneering countries,
more senior member, and there are opportunities efforts that would lay the groundwork for our disciplines,
to expand this initiative. This concept, expanding field. Similarly, Philip Morse – who was Blackett’s
Mentor Match to serve more early-career members, friend and colleague – organized a group of generations &
was a recommendation from the 2024 Ad Hoc scientists to assist the U.S. Navy during the
Committee on Early-Career Professionals. war [4]. The origins of O.R. owe their success
perspectives –
Equally important is the support for peer to the power of collaboration, demonstrating advancing
collaborations, enabling members – whether the potential that can be harnessed by bringing
they are across the globe or just across town – together unique perspectives from across the
smarter
who are facing similar challenges to share ideas, globe, different generations and a variety of decisions for a
experiences and solutions or, at the very least, to backgrounds to work toward common goals.
find a sympathetic ear. Another recommendation better world.
from the 2024 Ad Hoc Committee on Early-Career Conclusion
Professionals was to expand local and regional These ideas are not new, but their importance
meetings to create more opportunities for has grown significantly in the current political
collaboration, especially for early-career climate. My focus remains on making informed
members who may be unable to secure decisions on behalf of INFORMS that best
funding or spare the time to attend support the interests of our members and the
national meetings. general public. By fostering a strong community
and encouraging collaboration across countries
Collaboration Between and cultures, disciplines, generations and
Academia and Industry perspectives, INFORMS strengthens not only
With ongoing uncertainty surrounding U.S. federal our profession but also the communities we
funding, the ability to bring together industry and serve – advancing smarter decisions for a
academia to tackle real-world, pressing problems is better world.
fundamental to the mission of INFORMS. However,
this collaboration is only effective if we maintain a Note. The reference list can be found online:
good balance of academic and industry members https://doi.org/10.1287/orms.2025.02.05.
within INFORMS. Expanding local and regional
meetings can help maintain a good balance and DAVE HUNT is a vice president at the management consulting firm
build these collaborations. Oliver Wyman. He is the 2025 INFORMS president.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 7

HQ HIGHLIGHTS

FINDING COMMON
GROUND
We may disagree on the
“how” but not the “what.”

BY ELENA GERSTMANN
AND ASHLEY SMITH

THIS COLUMN WILL BE A LITTLE DIFFERENT, act on them. It’s also about how, even when we
and that’s intentional. For years, Ashley Smith, come from different perspectives, we often want
INFORMS Public Affairs Coordinator, has worked the same thing: fairness, opportunity, respect and
behind the scenes to help Elena write these the chance for everyone to contribute.
columns – a common practice in public affairs.
But for this one, we felt it was time to speak The “Why” We Can Agree On
openly together. Why? Because we have different Our experiences shape how we think about
political leanings, but we are aligned on goals. fairness and access. But when we step back
With Juneteenth Although politics and personal history don’t define and talk to each other – not around each
commemorated excellence, they can shape how we understand other – we often realize we’re aiming for the
challenges, build relationships and help our same outcomes.
this month, many community thrive. That’s why we believe it’s We want members to succeed based on the
important to move beyond communicating in strength of their ideas and contributions – not
organizations, shorthand – like using the term “DEI” when we based on where they’re from, what they look like
including might disagree on its meaning – and focus instead or who they know. We want people to feel a sense
on what we want for INFORMS. On that, we agree. of belonging at our conferences, in our journals,
INFORMS, are Why now? Because June marks a time of on our boards and throughout our community.
taking time to reflection and conversation across the country We want INFORMS to be a professional home
about freedom, justice, security and equitable for everyone who is passionate about operations
ask what it truly opportunity. With Juneteenth commemorated this research and analytics.
means to be a month, many organizations, including INFORMS, The “how” will vary. Some members work to
are taking time to ask what it truly means to be a improve representation. Others mentor early-
welcoming and welcoming and fair community. For us, that means career professionals or advocate for change
going beyond statements – it means building within their institutions. These approaches may
fair community. something lasting and real. look different, but they’re all pushing toward the
This column reflects both of our voices, same goal: a stronger, more welcoming profession.
our experiences and our genuine collaboration.
It mirrors something we both deeply believe Expanding Opportunity
about INFORMS: Bringing together different Strengthens Our Work
perspectives makes our message stronger and Let’s say it clearly: This is not about lowering
our community better. It’s about the values we the bar. It’s about making sure the bar is visible,
share as a professional community and how we reachable and fairly set. When more people are

8 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

included, we raise the quality of our conversations core values that matter. When we listen – really
and the caliber of our work. We uncover new listen – we often find we have more in common
insights, ask better questions and tackle than we think.
bigger problems. These moments remind us that our differences
The work of our members is also more important don’t have to divide us. They can help us grow.
than ever. The landscape in Washington, D.C.,
is shifting – programs and policies that once What INFORMS Can Be
supported equity in research, science and Imagine an INFORMS in which everyone –
education are being revisited or rolled back. regardless of background, identity or point of
Across the country, institutions are grappling view – feels like they belong. Where our programs
with how to ensure fairness while navigating and publications reflect the full range of talent in our
legal, political and cultural pressures. These profession. Where first-time attendees and longtime
are not abstract debates; they impact funding, leaders alike see a future for themselves here. Some
access and the future pipeline of talent in say we are already here, while others think we
our field. still have work to do – but neither of us has met a
At INFORMS, we’ve seen how expanding member who disagrees that this is our North Star.
opportunity leads to better outcomes. Opening What matters most is understanding the journey Our
leadership and editorial roles has strengthened never ends. We will never have it completely figured
our publications. Member-led forums have created out or be able to check that box as “done.” It’s about experiences
space for connection and collaboration that didn’t continuously getting better and moving the needle, shape how we
previously exist. And our recent efforts to make and we are. Deliberate work focused on creating
conferences more accessible are helping us better a more inclusive INFORMS has made an think about
understand and engage our full community. impact, including: fairness and
None of this would be possible without members • Nearly half of INFORMS journal editors-in-chief
who step up, speak out and shape the future of our are now women. access. But
organization. These aren’t top-down mandates; • Captioning and accessibility tools are being
they’re member-driven movements. piloted to improve the conference experience
when we step
for all attendees. back and talk to
Personal Stories, Shared Purpose • New forums like the Military Veterans Interest
We don’t share our personal perspectives to make Forum and the Pride Forum are expanding each other –
this about us. We share them because they reflect a our sense of what professional support can not around
broader truth: Finding common ground is possible, look like.
and it makes us better. each other –
Elena’s family history is a story of resilience. Her When we ensure a more welcoming INFORMS, we often realize
grandfather escaped Austria during World War II, we all benefit. Think of wheelchair ramps: Originally
hidden in a bakery truck to escape persecution. designed for a specific need, they now help parents we’re aiming
Her father and grandmother were later able to with strollers, travelers with rolling suitcases and
join him in Britain, where a family took them in delivery workers. The same is true here. When we
for the same
and gave them shelter. Their survival depended design systems that are open and fair, we make the outcomes.
on people who acted when it mattered most – entire profession stronger.
people who found ways to help rather than simply
reacting to the world around them. That story One Community, Many Voices
informs her view that real progress comes from We are all stewards of this organization, and we are
helping each other, especially when it’s hard. stronger when we move forward together – not just
Ashley’s family history can be traced to the in our technical contributions, but in how we treat
dawn of the American revolution with a direct each other and shape the culture of our field.
line to relatives who fought and persevered to give This isn’t about checking a box. It’s about opening
birth to our country. Yet her personal story is about doors and making space for more voices, more
learning and growing through conversation. In perspectives and more breakthroughs. At the end
2020, during the national reckoning that followed of the day, we’re all on the same page – even if we
the murder of George Floyd, a close family member – got here by different routes. And that’s exactly how
who holds deeply conservative views – expressed progress is made.
strong opposition to the protests and public outcry.
But when asked, “Do you think it’s okay for Black ELENA GERSTMANN, Ph.D., FASAE, CAE ([email protected]),
people to be hurt by police?” their answer came is the executive director of INFORMS (5521 Research Park Drive,
quickly: “Of course not.” That moment revealed Ste. 200, Catonsville, MD 21228). She can be reached via email.
something powerful. We may disagree on how to
approach an issue, but we can still agree on the ASHLEY SMITH is the public affairs coordinator at INFORMS.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 9

INFORMS AWARDS &
PRIZES DEADLINES
George B. Dantzig Dissertation Award June 30, 2025

Prize for the Teaching of OR/MS Practice June 30, 2025

Saul Gass Expository Writing Award June 30, 2025

Undergraduate Operations Research Prize June 30, 2025

Volunteer Service Award June 30, 2025

Donald P. Gaver, Jr. Early Career Award July 7, 2025

INFORMS Case Competition July 31, 2025

Franz Edelman Award Oct. 1, 2025

UPS George D. Smith Prize Nov. 24, 2025

INFORMS Early Career Practitioner Award Dec. 1, 2025

INFORMS Prize Dec. 15, 2025

Impact Prize Apr. 30, 2026

For more information and to submit an application to the above

awards, please visit https://www.informs.org/Recognizing-
Excellence/INFORMS-Prizes.
TOP SOCIAL POSTS
1 2 3 4 Check out
Founding Director of INFORMS Member 2025 Edelman Winner Tribute to John D.C. the top 8
Georgia Tech's MS in in a Minute with at the 2025 INFORMS Little, a giant in O.R. and most-clicked
Analytics program, Joel Ahmad Bassaleh Analytics+ Conference: marketing science. Learn
Sokol, shares how the INFORMS Member USA Cycling Awarded the about his contributions items on
program began — and how in a Minute 2025 INFORMS Edelman to the field, explore INFORMS'
it has grown into a finalist Award for Data-Driven critical links to historical
for the prestigious UPS Olympic Gold Strategy documentation & draw social media
George D. Smith Prize. INFORMS YouTube out his lessons. channels
INFORMS YouTube OR/MS Today
last month.
Connect with us
5 6 7 8 on your favorite
The analytics field has INFORMS is proud The INFORMS Industry- This seminar describes how
evolved. CAP-Expert (CAP- to recognize long- Academia Collaboration LLMs are being used at
channel!
X) – the new pinnacle time member and Committee (CIAC) is Microsoft to improve supply
certification for analytics distinguished scholar soliciting articles on chain efficiency.
leaders who drive James J. Cochran on insights, successes & INFORMS YouTube
organizational success. his election as a Fellow failures learned from
CAP of the African Academy industry-academia
of Sciences (AAS). collaborations in O.R.,
OR/MS Today advanced analytics &
related fields.
INFORMS CIAC

RECENT RESOUNDINGLY HUMAN

PRESS RELEASES March

@INFORMS
We kicked off our Resoundingly Human podcast
Masculinity Over Money? The Hidden Barrier Keeping coverage of the 2025 Franz Edelman Award
Men Out of Top Occupations Competition – the Nobel Prize of Analytics – with
episodes featuring the finalist teams from SF Express,
New AI-Driven Model Could Revolutionize Lufthansa, and USA Cycling. They shared unique
Conservation Efforts for Endangered Species insights into their projects, which represent the most
impactful analytics applications from around the world.
AI-Powered Tech Supercharges Ocean Cleanup,
Boosting Plastic Collection by 60% April
Our Franz Edelman Award Competition coverage
@INFORMS_orms

MEMBERS IN THE NEWS continued with interviews featuring the finalist teams
from Amazon and WM. With the competition occurring
in April, these episodes captured the excitement of
An American-Made iPhone: Just Expensive or the finalists before sharing their impressive projects
Completely Impossible, The Wall Street Journal, with the INFORMS community.
featuring Tinglong Dai
May
What Trump’s sweeping tariffs mean for the US economy
This month featured an interview with tech
and you, ABC News, featuring Rob Handfield entrepreneur Matthew Brady, recorded in-person
Buy now or wait? What to know as tariffs loom on during the 2025 INFORMS Analytics+ Conference,
food, tech, apparel, The Washington Post, featuring that explored a new app that helps students navigate
Sunderesh Heragu important decisions. Then we wrapped up our 2025
Franz Edelman Award coverage with an interview
In 2025, you can’t have an effective democracy without with the finalist team from Flipkart, who shared their
data literacy, Federal News Network, by Jeff Camm perspective on what it was like to compete this year.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 11

ADVOCACY IN D.C.

ADVOCACY IN A TIME OF TURBULENCE

BY JEFF COHEN

WASHINGTON, D.C., IS FACING A But advocacy isn’t, and shouldn’t be, just about
convergence of pressures: fiscal constraints, what INFORMS is doing directly. INFORMS is
rising global complexity, and a policy environment readying the launch of a new INFORMS Advocacy
reshaping how science and innovation are valued Toolkit to better equip members with messaging,
and supported. These shifts are rippling across meeting guides and outreach strategies to help tell
the country and are especially critical regardless your story and call attention to the ramifications
of whether you work in academia, industry of policy discussions at all levels of government
or government. by sharing how your work translates into real-
You’ve seen the headlines warning that federal world impact. Look for the announcement
science funding is at risk. Congress continues to and information about how to access the
govern through short-term continuing resolutions toolkit on INFORMS Connect, or check
while proposals to cut nondefense discretionary this article on OR/MS Today online for a
spending – including National Science Foundation, link to the toolkit [1].
National Institutes of Health and Department of In the meantime, here are concrete ways to
Education budgets – grow louder. Even programs engage in advocacy.
aligned with national priorities are being scrutinized
through new ideological and economic lenses, Speak in Plain Language:
including capping indirect costs and tying research Engaging Congress
funding to institutional civic behavior. Members of Congress don’t wake up thinking
Yet, Washington is still paying close attention to about stochastic optimization, queuing theory, the
science, especially in artificial intelligence (AI), data knapsack problem or any other specific aspects of
science and algorithmic governance. The White what you do or the models you use and create. But
House is advancing a new three-part innovation they do care about improving healthcare delivery,
agenda to modernize the U.S. science ecosystem reducing traffic congestion, securing elections,
and strengthen America’s leadership in emerging ensuring national and economic security, and
technologies. Meanwhile, Congress is advancing lowering costs for their constituents – this is
various legislative frameworks, the administration your connection.
is issuing executive orders and bipartisan efforts to Writing another white paper doesn’t generally
support STEM continue to move forward. help them. It’s another in a growing stack of
A case in point: the recent unanimous House materials that the typical person working in a
passage of the Mathematical and Statistical congressional office won’t ever read. What they
Modeling Education Act (H.R. 730), in which really need is a person. A trusted local expert
INFORMS successfully worked to embed real- who can explain or, better yet, bring to life how
world modeling, operations research and analytics analytical models, data-driven decision tools and
into K-12 classrooms. This win is a prime example systems thinking are driving actual outcomes in
of proactive advocacy translating into a strategic their district or state.
road map for preparing the next generation of data- That person should be you.
literate problem-solvers. It explicitly reinforces Action step: Reach out to your representative’s
the importance of operations research, analytics office. Schedule a meeting – virtually or in person.
and data in meeting the long-term needs of the Invite them to see your lab, classroom or project
United States. in action. Bring the data but lead with the human
Through all of this, INFORMS has been impact. If you’re unsure where to start, please reach
actively engaging congressional offices to brief out to me and I’d be glad to help.
staff, highlight member work and connect our
field to real policy priorities. We work diligently Visibility is Not Vanity: Public Relations
and in a nonpartisan way to educate the policy If people don’t know your work exists, it doesn’t
community about INFORMS, our member matter how brilliant it is. To be clear, I don’t
experts, and how our fields of interest contribute mean what your academic colleagues see in
to stronger policymaking and government peer-reviewed journals – I mean how your
operations. Similarly, we work with the media work translates into the media read and valued
to ensure that our members and their expertise by policymakers, business leaders and other
are visible to national decision-makers. influencers. This means turning complex models

12 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

and areas of expertise into relatable narratives and advantage of INFORMS’ resources. You don’t have
publishing impact stories in places decision-makers to try and change the world – you just need to try
read and value. and make a difference.
INFORMS has had a lot of success with this.
But INFORMS can only do so much on its own. We Final Thoughts: Advocacy is
need more members stepping forward and building More Essential than Ever
public awareness at the grassroots level. What we’re seeing in Washington is a signal
Action step: In addition to INFORMS, use your of a longer-term shift. The status quo of how
institution’s communications teams to share your Washington advances and supports science is
work and expertise. Don’t wait for them to come changing. Universities and researchers are under
to you – be proactive. Write opinion editorials (op- heavier scrutiny. Funding priorities are shifting. And
eds). Teach them about your work and why they the value of the federal government’s investment
should devote time and effort to promoting it in the in science is increasingly being judged through
media. Translate your results into plain language subjective lenses of public return, national interest
and post it to social media. INFORMS Advocacy and the role of government.
Toolkit is now available, use it to sharpen That’s not going away next month, next year or
your messaging and push your work into the even after the next election cycle. These shifts are
right conversations [1]. part of a fundamental structural realignment that
will shape how research is funded, how expertise is
Contribute Your Expertise: treated and how our work is viewed by the public
Relationships Build Influence and policymakers for years to come.
You don’t need to be a lobbyist to influence policy. That means we have both short- and long-
You just need to be a dot-connector who builds term imperatives.
bridges between ideas, communities and outcomes. In the short term, we must be visible. Speak
When you speak at community forums, contribute up. Be constructive and positive. Show how our
to advisory boards or brief local decision-makers, you fields – and your work – are improving national
do more than share research; you build trust. And security, economic resilience, public health and
trust is the currency that moves ideas from concept infrastructure, not in theory but in practice.
to implementation. In the long term, we must continue building trust
Action step: Find local opportunities to and reinforcing relevance. That means developing
provide your expertise. Serve on a committee. stronger relationships with policymakers, aligning
Partner with nonprofits. Let people see what with public priorities and ensuring our fields are
operations research, analytics, data science, academically respected and publicly indispensable.
decision models and more look like in action. We can’t do that by staying quiet or siloed. We do it
Over time, these local relationships grow into by engaging – consistently, credibly and strategically.
national influence and bolster a strong public At INFORMS, we focus on doing just that – and
reputation for our field. we’re counting on you to join us. The success of H.R.
730 shows what’s possible when we connect our
Start Where You Are – But Start! work to national needs. But we can’t stop there. If
Many people think advocacy is someone else’s we don’t define the value of our fields, someone
job. And yes, it is my job. But it’s your job, too. else will. And we may not like what they say.
When you stay silent, influence is ceded to This is our moment to lead the conversation –
those who speak the loudest, not those with not just about O. R., analytics or data science but
the deepest understanding. about why science matters and why Washington
If we don’t define and reinforce the value of our should invest in it wisely. Let’s make it count.
fields to society, someone else will – and we may
not like their version. Author’s note. Email Jeff Cohen ( [email protected])
Action step: Do one thing this month to get for help in getting involved in INFORMS advocacy efforts.
involved. Join the INFORMS advocacy effort,
reach out to your local elected official or post a JEFF COHEN is chief strategy officer at INFORMS, where he leads the
organization’s advocacy and public affairs efforts. Connect with him at
short social media thread about your work. [email protected] or linkedin.com/in/jmcsc to learn more about how
Contact us to learn how to plug in. you can become involved in advocacy.
The good news is that you don’t need to do
it all at once. It’s okay to take small steps. But REFERENCE
start somewhere. Find your comfort zone. Take 1. https://bit.ly/4kRwitT

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 13

FORUM

THE OPPORTUNITY THAT LIES IN IoT

DATA FOR ANALYTICS ENTREPRENEURS
BY KEN MISEDA

MY FASCINATION WITH THE PROLIFERATION deep curiosity in me: If a single IoT device, such as
of the Internet of Things (IoT) began during the a smartwatch, could create an entire ecosystem,
If a single IoT COVID-19 pandemic. Like many people, I spent a what other possibilities exist within the expanding
device, such as significant amount of time indoors and decided realm of IoT?
to take up running to stay fit. Initially, I thought The origins of IoT trace back to 1989, when David
a smartwatch, running was just about getting a good pair of Nichols and his colleagues at the Massachusetts
could create shoes and comfortable workout clothes. However, Institute of Technology (MIT) created the first IoT
I soon discovered a vast ecosystem built around device [1]; a key milestone was reached in 2008
an entire this simple activity. when the number of connected devices surpassed
ecosystem, My phone automatically tracked my steps, the global population [2]. The following year saw
providing daily updates on my activity levels, the advent of Bluetooth 3.0 [3], which enabled
what other distance covered and step count. This piqued my faster data transfer speeds of up to 20 megabits
possibilities interest, leading me to invest in a smartwatch.
Through this device, I was introduced to online
per second (Mbps), accelerating IoT adoption.
Several advancements over the years contributed
exist within platforms that helped manage running sessions. to the IoT boom, including the development of low-
What surprised me the most was the sheer number power wide-area network (LPWAN) technology [4],
the expanding of people using these platforms – millions of reduced production costs of industrial IoT sensors [5]
realm of IoT? smartwatch users syncing their data to online and the rise of IoT platforms. Between 2015 and
databases in real time. This realization sparked a 2017, major technology companies entered the IoT

14 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

space – Amazon launched Amazon Web Services Despite the immense potential of IoT, challenges
(AWS) IoT Core, Microsoft introduced Azure remain, particularly regarding data security and
IoT Hub and Google unveiled Google IoT Core. ethics. Most IoT-generated data is user-driven [8],
These platforms paved the way for widespread often containing personal, health or sensitive
IoT adoption. By 2021, the number of connected information. This raises significant privacy
devices exceeded nonconnected devices, and concerns. A notable example is the March 2024
by 2022, IoT ranked among the top three most New York Times report, which revealed that
impactful technologies according to the World companies such as GM, OnStar, LexisNexis and
Economic Forum [6]. Estimates suggest that by Verisk Analytics had collected and sold driver
2025, the economic impact of IoT could range behavior data to insurance companies without
from $2.7 trillion to $6.2 trillion per year. user consent [9].
The expansion of IoT devices has led to an While such challenges exist, they do not
explosion in data generation. IoT devices are overshadow the opportunities within IoT and
equipped with sensors that continuously collect data analytics. According to IoT Analytics, between
information from everyday objects, from household 2016 and 2024, the top 20 IoT startups collectively
appliances to industrial machinery. According to a received nearly $6.7 billion in funding [10], with
2020 forecast by the International Data Corporation investment expected to grow. This demonstrates
(IDC), the number of IoT devices is expected to increasing support for IoT-driven innovations.
reach 41.6 billion by 2025, generating an The rapid evolution of IoT presents both
astonishing 79.4 zettabytes (ZB) of data [7]. challenges and opportunities. As devices
Data is the foundation of the IoT ecosystem. become more interconnected and data As an analytics
The real-time insights provided by IoT devices
enable businesses to make data-driven decisions,
generation increases, the need for ethical data professional, I
handling, security measures and regulatory
optimize operations and improve efficiency. Even at frameworks will be crucial and is an opportunity see the future
a basic level, leveraging descriptive and diagnostic
analytics allows companies to significantly enhance
in itself. At the same time, the entrepreneurial
potential within IoT is vast. By responsibly
of IoT not only
their products and services. harnessing IoT data, analytics professionals can as a vast data
The integration of artificial intelligence (AI) with drive innovation and create impactful solutions in
IoT further amplifies data analytics capabilities. this ever-expanding digital ecosystem.
landscape
AI algorithms have become more sophisticated,
KEN MISEDA is an experienced business data analyst with over
but also as
enabling organizations to analyze vast and complex
datasets with greater precision. This fusion of seven years of expertise in data collection, analysis and visualization. a realm of
He holds an M.S. in business analytics and has a proven track record
AI and IoT allows businesses to uncover subtle
patterns and trends that would have otherwise
of developing processes and conducting advanced statistical analyses entrepreneurial
to optimize performance and enhance operational efficiency.
gone unnoticed. opportunity.
Returning to my personal journey with running, REFERENCES
I became intrigued by the businesses that emerged 1. Salunke, V.M. and S. Mahajan, 2024, “The evolution and
implications of the Internet of Things,” March, DOI: 10.13140/
from the IoT data ecosystem. A notable example is
RG.2.2.16409.39527.
Strava, a fitness-tracking platform founded in 2009. 2. Evans, D., 2011, “The Internet of Things: How the next evolution
Strava records users’ activity data, offering insights of the internet is changing everything,” White Paper, Cisco.
such as route summaries, elevation changes, speed, 3. https://majorhifi.com/the-difference-between-bluetooth-
timing, power and heart rate. Beyond just tracking 3-0-4-0-4-1-4-2-explained/
4. https://www.techtarget.com/iotagenda/definition/LPWAN-low-
workouts, Strava integrates social networking
power-wide-area-network
features, allowing users to share their progress with 5. https://www.statista.com/statistics/682846/vr-tethered-hmd-
others. This illustrates how IoT data can be harnessed average-selling-price/
to create innovative and profitable enterprises. 6. Manyika, J., M. Chui, J. Bughin, R. Dobbs, P. Bisson and A. Marrs,
As an analytics professional, I see the future 2013, “Disruptive technologies: Advances that will transform life,
business, and the global economy,” McKinsey Global Institute, May.
of IoT not only as a vast data landscape but also 7. https://ieeexplore.ieee.org/document/9210375/references
as a realm of entrepreneurial opportunity. With 8. Pomella, K., 2024, “The impact of IoT on customer data collection
the anticipated growth of IoT devices and the and analysis,” September 16, https://kenpomella.com/blog/
increasing volume of data they generate, analytics the-impact-of-iot-on-customer-data-collection-and-analysis.
professionals are positioned to become not 9. https://www.nytimes.com/2024/03/11/technology/carmakers-
driver-tracking-insurance.html
just data consumers but also innovators, 10. Kaman, Z., 2024, “IoT startup landscape 2024: 7 notable
developing solutions embedded within the insights,” IoT Analytics, July 31, https://iot-analytics.com/
IoT ecosystem. iot-startup-landscape.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 15

VIEWPOINT

AI-DRIVEN REMODELING IN HEALTH

PLAN WORKFLOWS: REVOLUTIONIZING
HEALTHCARE ANALYTICS
BY AISHWARYA AIREN

ARTIFICIAL INTELLIGENCE (AI) HAS A • Changes to social determinants of health:

powerful impact on the healthcare industry With patient lifestyle changes, we need to
because it can pivotally impact and redefine capture and effectively track those adapting
traditional healthcare workflows and analytics. metrics to enable more accurate assessments.
From managing effective patient flows and • Streamlined authorization reviews: During
combatting compromised surge, to providing authorization processes, AI can summarize
correct authorization requests and relevant relevant notes, provide citations and suggest
service authorizations, to identifying population additional data points, such as past encounters
of patients for value-based care and program or related authorizations. This reduces the
outreach, data modeling is everywhere. With time staff spend searching for information
extensive experience across various healthcare and ensures accuracy.
modules, I see immense potential for AI-powered
tools to transform health plan workflows and These are just some of many examples
enhance efficiency. The possibilities include using of AI-driven features that are essential for
AI to optimize operational workflows, identify cost- organizations managing large volumes of data
saving opportunities and improve decision-making while striving to maintain operational efficiency.
through predictive analytics.
Optimizing Population Health Management
Elevating Efficiency with AI-Assisted and Outreach
Health Plan Workflows Identification of a specific patient population to
One of the most crucial tasks in healthcare is to drive better care and address care gaps is essential
navigate a specific workflow within a module in before planning to enroll them into a program. The
the quickest time possible. Because this is time- role of AI in population health analytics is equally
intensive, with AI tools and their faster processing transformative. Extracting population-specific
ability, we look forward to groundbreaking features data from a large dataset through generative AI
to address some of the most time-intensive tasks in integration would allow nontechnical users to
health plan management. generate reports, which can then be leveraged
• Interactive dashboard insights: When to outreach to participants in a timely yet
providers land a dashboard relevant to their effective pattern.
practice, AI-generated summaries can allow
users to quickly interpret key takeaways from Tailoring Patient-Provider Messaging
dashboards. By conversationally describing In addition, AI assists in crafting appropriate
insights, interpreting data and suggesting next quick input messages. By tailoring content based
steps, this feature minimizes the need for deep on user input for tone and style, AI streamlines
dives into raw data, making decision-making communication while ensuring personalization.
faster and more informed. This capability empowers healthcare organizations
• Recent drug-interaction warnings: One of the to effectively engage with diverse populations,
essential metrics for patient care and safety enhancing the impact of outreach efforts.
is to review the most common prescribing
practices to determine how often patients are Image-to-Text Automation
prescribed medications that can interfere with Like with most smartphones, AI can help convert
their existing medications. Getting up-to-date vendor or externally generated managed care
alerts of the number of recent drug interactions coverages existing in a scanned document format
can help with improvement practices. to text. It can also leverage the ability to prefill

16 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

relevant form fields with standardized naming the focus should be on strategically adopting these
conventions upon a quick scan. These AI-powered technologies and sharing insights to shape the
solutions convert scanned documents into system- future of healthcare [3].
readable data, significantly reducing staff workload
while minimizing errors. Editor’s note. This article first appeared as a
blog post on Women in Analytics (https://www.
The Bigger Picture: AI’s Role in the womeninanalytics.com/).
Future of Healthcare Analytics
AISHWARYA AIREN is a highly skilled and results-driven business
The integration of AI into healthcare signals a new
era of efficiency and innovation. By automating
intelligence developer with a proven track record in healthcare The integration
analytics and advanced technologies. With a robust background in
repetitive processes, summarizing complex data engineering and big data tools, she has been instrumental in of AI into
datasets and providing actionable insights, these leading data-driven initiatives that empower healthcare organizations
tools empower healthcare organizations to to make informed decisions. Aishwarya’s expertise spans Epic systems healthcare
applications, HL7/FHIR integration, Power BI and full-stack data analysis,
transition toward value-based care models while
showcasing her ability to bridge technical acumen with strategic vision. signals a new
addressing the operational challenges of an ever- She is currently focused on leveraging artificial intelligence in healthcare
growing healthcare landscape [1]. analytics to elevate patient care, streamline workflows and shape the era of efficiency
future of the healthcare industry.
A Call to Action
and innovation.
REFERENCES
The time to embrace AI in healthcare analytics is
1. World Economic Forum, 2020, “AI and Healthcare: A
now. Studies show that AI can improve operational Perspective,” https://www.weforum.org/whitepapers/
efficiency by up to 30% and reduce healthcare ai-and-healthcare-a-perspective.
costs by an estimated $150 billion annually by 2. Forbes Insights, 2019, “AI and Healthcare: A Giant Opportunity,”
2026 [2]. AI-embedded tools have the potential February 11, https://www.forbes.com/sites/insights-
to revolutionize workflows, optimize resource intelai/2019/02/11/ai-and-healthcare-a-giant-opportunity/.
3. Fortune Business Insights, 2023, “Artificial Intelligence in
allocation and enhance patient care quality, Healthcare Market Size, Share & Trends Analysis Report, 2023-
ultimately driving significant return on investment. 2030,” https://www.grandviewresearch.com/industry-analysis/
For healthcare professionals and organizations, artificial-intelligence-ai-healthcare-market.

INFORMS NavigatORs
are back!
Guiding #INFORMS2025 Attendees!
INFORMS is excited to continue the NavigatOR volunteer
opportunity for the 2025 INFORMS Annual Meeting in
Atlanta, Georgia. This group of members helps promote
community, connection, and knowledge-sharing at the
Annual Meeting, as well as building a vibrant social media
and in-person presence for INFORMS.

Look for communications soon on how to apply to become

a NavigatOR!
W E
*If you attended the 2024 INFORMS Annual Meeting,
you may remember this helpful group of volunteers! S
https://info.informs.org/informs-navigators
PROFESSOR PADMA P. PATHFINDER
LETTER TO SELF

REFLECTIONS ON TRANSITION
TO CIVILIAN LIFE
A Letter to Myself a Year Ago

BY SCOTT COHICK, CAP-X

Dear Scott, familiarized yourself with LinkedIn and tackled

A year from now, you will be writing this letter, tasks head-on that are often unfamiliar to career
reflecting on your transition out of the Navy and military personnel (i.e., writing a compelling
which actions had the greatest impact on this time resume, job interviews, employment negotiations).
in your life. You will discover a passion in assisting You leveraged your network and numerous
others in understanding the steps they can take to transition support organizations to learn and
help them navigate turbulent times. Spoiler alert: practice these unfamiliar and required skills.
The good news is that you’ll find a new professional Each action was a deliberate investment in your
home – one that challenges you intellectually future, equipping you with the confidence and
and aligns with your values. You will successfully readiness to step into the civilian workforce.
navigate this transition – not without effort, but Your approach to decision-making was another
with vision and determination. And when you look strength. You didn’t view this as a simple job search –
back, you’ll realize that the most important steps you understood it was about creating a life that
were the ones you took well before your military aligned with your priorities. You approached
retirement date. your transition like the analyst you are – framing
In the year leading up to June 2024, you laid the the problem, gathering information, selecting
groundwork for a successful transition by focusing your approach, iterating and refining, and finally,
on self-reflection, preparation and networking. One deploying yourself. Family considerations, work
of the most impactful decisions you made was to flexibility, professional growth and location all
start early. You didn’t wait until retirement to start factored into your analysis. By methodically
planning your next steps; instead, you took the time identifying your critical variables and constraints
to ask yourself critical questions: What motivates and considering trade-offs, you put in motion your
me? What kind of work brings me satisfaction? transition plan with confidence and determination.
What work-life balance does my family need? These
reflections revealed not only your ideal career and "
employer options but also the life you envisioned
for yourself and your family. The path you took wasn’t always
You also knew better than to transition alone. easy, but it was intentional.
When faced with something as challenging as
stepping into a completely new professional
world, you recognized that navigating it alone Preparation and support of your network resulted
would be unwise. Seeking assistance, asking for in a transformative transition. You were welcomed
advice and learning from others’ experiences into the role of a principal operations research
became invaluable. The personal and professional analyst at CANA, LLC, where you now apply your
relationships you cultivated over the years – expertise in operations research to national defense-
particularly through the INFORMS community – related projects. The supportive and collaborative
became a vital part of your transition journey. culture at CANA aligns with your values and provides
Networking within INFORMS not only connected the challenges and balance you sought. Your journey
you with mentors and colleagues who shared their demonstrated that a proactive, thoughtful approach
wisdom but also introduced you to opportunities to transition can yield outcomes far beyond simply
that helped shape your path forward. The insights, finding a job – it can open the door to a meaningful
encouragement and camaraderie within this and rewarding path forward.
society reinforced your confidence in stepping As you reflect on your transition, gratitude
into the next phase of your career. takes center stage. You’ll remember the mentors
In addition to leveraging your network, you who shared their wisdom, the colleagues who
committed yourself to learning the tools and offered recommendations and the organizations
skills necessary for a successful transition. You that provided resources. You’ll appreciate the

18 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

network of people who made this journey possible
and the effort you put into building connections.
And when you decide to begin paying it forward,
you will realize that you are now in a position to
share your experiences and encourage others to
approach their transitions with the same vision
and determination.
The path you took wasn’t always easy, but it
was intentional. A year from now, you’ll look
back on this transition not with fear or frustration
but with pride and a sense of accomplishment.
You’ll know that every step you took – from
reflection and preparation to relationship-
building and decision-making – helped shape
the next chapter of your life.
So, Scott, keep moving forward. Stay curious
and adaptable, and don’t hesitate to lean on the
support of those around you. The road ahead is
bright, and you have all the tools to navigate it
with purpose and success.
Sincerely,
Your Future Self
SCOTT COHICK, CAP-X, is a principal operations research analyst at CANA,
LLC, bringing more than 23 years of experience as a Naval Aviator and
defense analyst. He concluded his Navy career in 2024 as military faculty in
the Operations Research Department at the U.S. Naval Postgraduate School.
His expertise spans wireless power transmission modeling, resilience
modeling, simulation and advanced manufacturing. Scott holds a master’s
degree in operations research from the Naval Postgraduate School and
serves as vice president of the INFORMS Military Veterans Interest Forum.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 19

QUANTUM COMPUTING HAS THE Despite growing interest, demonstrating
potential to transform many areas of computing, a clear quantum advantage for optimization
including – possibly – operations research remains an open challenge. Benchmarks often
(O.R.). The United Nations declared 2025 the show that classical heuristics still outperform
International Year of Quantum Science and near-term quantum algorithms on practical
Technology, suggesting that this is a good time problems, especially when accounting for noise
for INFORMS members to ask themselves if they and circuit depth constraints. As such, quantum
are interested in learning about, contributing to optimization research is increasingly focused on
or using this technology. This article overviews problem instances with a specific structure – e.g.,
some of the areas of quantum computing that low treewidth, high degeneracy or symmetry –
potentially overlap with operations research; that may favor quantum heuristics. Continued
while we do not aim to (and could not possibly) progress will depend on tighter integration
be exhaustive, we provide a few pointers to between quantum algorithm design and classical
facilitate further investigation. optimization insights, particularly from fields such
as discrete optimization, global optimization and
Optimization computational complexity.
Several potential use cases of quantum
computing that have captured the attention of Simulation
researchers fall within the realm of optimization; The simulation of quantum mechanical systems
for a detailed discussion on the current state of is the application for which quantum computers
quantum optimization research, see [1]. Quantum were initially proposed by Richard Feynman in
optimization algorithms can be broadly categorized the 1980s. Mathematically, the evolution of such
by the type of hardware they target – near-term a system follows the Schrodinger equation. In the
or fault-tolerant – and by their solution strategy: simplest case – i.e., when the operator describing
exact or approximate. For near-term devices, the energy of the system does not vary over time –
variational quantum algorithms such as the the solution in closed form of the Schrodinger
Quantum Approximate Optimization Algorithm equation is a matrix exponential e-iHt applied to
(QAOA) aim to find approximate solutions to the quantum state, where H is a matrix and t is a
combinatorial problems using parameterized scalar. There are many quantum algorithms to
quantum circuits, although their empirical compute matrix exponentials of that type, given a
advantage remains unclear. For fault-tolerant description of H. This result underpins many other
devices, algorithms such as quantum interior quantum algorithms, including celebrated results
point methods and quantum semidefinite in quantum linear algebra, such as the Harrow-
optimization solvers offer theoretical speedup Hassidim-Lloyd (HHL) algorithm for linear systems
guarantees under specific assumptions that may of equations. Another application of this result is
not hold in practice. Fault-tolerant approximation the construction of Gibbs states: quantum states
algorithms with provable a priori performance described by a normalized matrix exponential.
guarantees have been suggested and in some Gibbs states directly generalize Gibbs distributions,
cases, they are shown to outperform (in theory) so under some conditions, it is possible to sample
the best known classical approximation algorithms from a Gibbs distribution more efficiently using a
for certain combinatorial optimization problems [2]. quantum computer than with a classical computer.
To demonstrate practical quantum advantage Manipulation of Gibbs states and sampling from
in optimization, systematic and rigorous Gibbs distribution are commonly found in quantum
benchmarking for challenging optimization optimization research. The intuition behind the
problems is key [3]. potential advantage of quantum computers in
Quantum optimization is not just about using this and related settings stems from the fact that an
quantum devices to solve classical problems; n×n matrix exponential e-iHt is only a (log)-qubit
optimization methods are also essential to make operator, so there can be considerable “space”
quantum algorithms and hardware more effective. advantage, leading to – in some cases – more
Classical optimization plays a key role in efficiently efficient algorithms.
compiling quantum circuits – reducing gate depth, Quantum computers may also be able to
optimizing qubit mappings or routing operations help with classical simulation tasks, such as
on hardware with limited connectivity. Linear and estimating probabilities by Monte Carlo and
integer programming methods are already used to related methods. The cornerstone of this potential
schedule quantum gates and minimize cross talk or advantage is amplitude estimation: a quantum
decoherence. Moreover, hybrid quantum-classical algorithm to estimate the probability of a given
schemes rely on classical solvers to train variational event to precision ϵ using O (1 ⁄ϵ) quantum samples
circuits, often using gradient-based or derivative- instead of O (1 ⁄ϵ 2) classical samples. Quantum
free methods. As devices grow more complex, samples are more powerful than classical samples,
optimization will remain central in the co-design so this is not a “plug-and-play” advantage but rather
of algorithms and hardware. points to an area with possible applications.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 21

QUANTUM COMPUTING

Classical simulation and uncertainty quantum resource demands. Financial

quantification are extremely useful for the optimization problems span continuous,
design and calibration of quantum hardware. discrete and mixed-variable formulations,
The hardware is affected by several types of noise, and quantum machine learning approaches
which can occur in various phases of the device’s face training and interpretability challenges
operation and are usually modeled based on similar to those in other fields [6].
physical principles. How to best calibrate the In chemical engineering, quantum computing
parameters of the noise models for a specific is being investigated for molecular simulations
hardware device and how to best use the relevant to catalysis and materials discovery.
information for optimal device control are These simulations feed into larger process
just two examples of areas in which INFORMS design and optimization frameworks and may
disciplines could be of great use for quantum be integrated with classical surrogate modeling
computer engineering. tools. While current hardware remains limited,
How to best hybrid approaches might be able to enhance
Machine Learning modeling of chemical systems [7].
calibrate the Quantum machine learning (QML) aims In transportation, a broad range of quantum
to leverage quantum algorithms to improve applications has been identified across road, rail,
parameters performance or efficiency in data-driven aviation, maritime and pipeline systems. These
of the noise tasks. Many QML algorithms mirror classical include predictive maintenance, congestion
counterparts, such as quantum support vector management, supply chain optimization,
models for machines, quantum principal component analysis emergency planning and simulation tasks such
a specific and quantum Boltzmann machines, but often as fuel efficiency, crash modeling and weather
assume idealized or fault-tolerant settings [4, 5]. forecasting. These use cases reflect core
hardware One often-cited motivation is representational OR/MS problems such as large-scale scheduling,
device and how efficiency: Quantum states over n qubits can network design and decision-making under
encode 2n amplitudes, enabling compact uncertainty; many are naturally suited for hybrid
to best use the representations of complex functions or quantum-classical implementations in digital
distributions. However, this does not twin environments [8].
information guarantee a speedup by itself. These examples reflect the broader potential for
for optimal On near-term devices, most QML methods OR/MS to guide the modeling, benchmarking and
rely on parameterized quantum circuits trained deployment of quantum computing tools. As the
device control via classical optimizers. These hybrid models field progresses, identifying problem structures
are just two face practical challenges such as barren plateaus, that match quantum capabilities will be key to
vanishing gradients and high estimator variance [4]. realizing impact. OR/MS expertise is also essential
examples of This creates a role for classical O.R. techniques – for building effective hybrid solutions that tackle
areas in which e.g., derivative-free optimization, robust training real-world complexity through collaboration with
or bilevel optimization – to improve reliability. quantum algorithm designers.
INFORMS Amplitude estimation (discussed earlier) also
Acknowledgments
disciplines provides a speedup that could benefit applications
such as risk assessment or Bayesian inference. In The INFORMS Quantum Computing and
could be of more engineering-focused tasks, O.R. tools can Operations Research Ad Hoc Committee was
help manage hybrid QML workflows by designing formed in October 2024. This article is written
great use for experiments, allocating compute resources or by its members.
quantum incorporating uncertainty. As entry points for
further reading, we refer to two recent reviews Disclaimer. This paper was prepared for
computer on the opportunities and limitations of QML [4, 5]. informational purposes with contributions from
engineering. the Global Technology Applied Research center
Applications of JPMorgan Chase. This paper is not a product
Quantum computing is being explored in multiple of the Research Department of JPMorgan Chase
domains in which OR/MS plays a central role, or its affiliates. Neither JPMorgan Chase nor any
including finance, chemical engineering and of its affiliates makes any explicit or implied
transportation. These areas provide structured representation or warranty and none of them
problems that are natural candidates for quantum accept any liability in connection with this paper,
techniques and serve as inspiration for other including, without limitation, with respect to
potential applications. the completeness, accuracy, or reliability of the
In finance, quantum algorithms are being information contained herein and the potential
developed for stochastic modeling, optimization legal, compliance, tax, or accounting effects
and machine learning. Techniques such as thereof. This document is not intended as
quantum Monte Carlo integration and gradient investment research or investment advice, or
estimation offer asymptotic speedups, even as a recommendation, offer, or solicitation for
though practical feasibility requires reduced the purchase or sale of any security, financial

22 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

instrument, financial product or service, or to be used TAMÁS TERLAKY is the Alcoa Endowed Chair Professor in the Industrial
in any way for evaluating the merits of participating and Systems Engineering Department, and director of the Quantum
in any transaction. Computing Optimization Lab at Lehigh University. He has published
four books, edited more than 10 books and journal special issues, and
DAVID E. BERNAL NEIRA is an assistant professor at the Davidson published over 225 research papers. Topics include theoretical and
School of Chemical Engineering at Purdue University, where he algorithmic foundations of math and quantum computing optimization.
and his Systems Engineering via Classical and Quantum Optimization He served as chair of INFORMS Optimization Society and the SIAM
for Industrial Applications (SECQUOIA) team focus on addressing AG Optimization. Terlaky’s recognitions include the Wagner Prize of
practical problems in science and engineering. His team has INFORMS; Fellow of INFORMS, SIAM, IFORS, The Fields Institute, and
specialized in discrete nonlinear optimization problems and Canadian Academy of Engineering. He is co-chair of INFORMS Quantum
Computing and Optimization Ad Hoc Committee.
applications of novel hardware, including quantum and physics-
inspired methods, to address them. Prior to joining Purdue University,
OR/MS
STEFAN WOERNER is a principal research scientist and manager of the
David was an associate scientist at the Research Institute for Advanced
Quantum Computational Science group of IBM Quantum at IBM Research
expertise is
Computer Science at the Universities Space Research Association
Europe – Zurich. He holds a Master of Science in Applied Mathematics and
and the Quantum and Artificial Intelligence Laboratory at NASA.
a Doctor of Sciences in Operations Management from ETH Zurich and is
also essential
Connect with him on LinkedIn at https://www.linkedin.com/
in/bernalde. renowned for his work in quantum finance, quantum optimization and
quantum machine learning.
for building
REBEKAH HERRMAN is an assistant professor of Industrial and effective hybrid
Systems Engineering at the University of Tennessee, Knoxville. Her REFERENCES
research interests lie in the intersection of combinatorics, discrete 1. Abbas, A., A. Ambainis, B. Augustino, A. Bartschi, H. Buhrman, solutions
optimization and quantum algorithms. C. Coffrin, et al., 2024, “Challenges and opportunities in quantum
optimization,” Nature Reviews Physics, pp. 1-18. that tackle
MOHAMMADHOSSEIN MOHAMMADISIAHROUDI is a Postdoctoral Jordan, S.P., N. Shutty, M. Wootters, A. Zalcman, A. Schmidhuber,
Research Associate in the Industrial and Systems Engineering
2.
R. King, et al., 2024, “Optimization by decoded quantum
real-world
Department at Lehigh University, working in the Quantum Computing interferometry,” arXiv: 2408.08292.
and Optimization Lab. He received his Ph.D. in industrial and systems 3. Koch, T., D.E. Bernal Neira, Y. Chen, G. Cortiana, D.J. Egger, R. Heese,
complexity
engineering from Lehigh University in 2024. His research
focuses on developing, analyzing and implementing efficient
et al., 2025, “Quantum Optimization Benchmark Library – The
Intractable Decathlon,” arXiv: 2504.03832, https://git.zib.de/qopt/
through
quantum and classical algorithms for solving large-scale
optimization problems.
qoblib-quantum-optimization-benchmarking-library. collaboration
4. Schuld, M. and N. Killoran, 2022, “Is quantum advantage the right
goal for quantum machine learning?,” PRX Quantum, Vol. 3, No. 3, with quantum
GIACOMO NANNICINI is an associate professor of Industrial and Art. no. 030101.
Systems Engineering, and Electrical and Computer Engineering at 5. Cerezo, M., G. Verdon, H-Y Huang, L. Cincio and P.J. Coles, 2022, algorithm
the University of Southern California. His main research interest is “Challenges and opportunities in quantum machine learning,”
optimization and its applications. Nature Computational Science, Vol. 2, No. 9, pp. 567-576. designers.
6. Herman, D., C. Googin, X. Liu, Y. Sun, A. Galda, I. Safro, et al., 2023,
RUSLAN SHAYDULIN is head of Quantum Engineering Research “Quantum computing for finance,” Nature Reviews Physics, Vol. 5,
at the Global Technology Applied Research center at JPMorgan Chase, No. 8, pp. 450-465.
where he and his team focus on practical aspects of evaluating quantum 7. Bernal, D.E., A. Ajagekar, S.M. Harwood, S.T. Stober, D. Trenev and F. You,
algorithmic speedups and realizing them on hardware. Areas of 2022, “Perspectives of quantum computing for chemical engineering,”
responsibility of Ruslan’s team include numerical benchmarking AIChE Journal, Vol. 68, No. 6, Art. no. e17651.
of quantum algorithms, compilation and execution on quantum 8. LeMaster, D.A. and P. Vakharia, 2024, “Quantum technologies
hardware, compilation to fault-tolerant architectures and error in transportation,” Office of the Assistant Secretary for Research
correction. Prior to joining JPMorgan Chase, Ruslan was a Maria and Technology, U.S. Department of Transportation, https://www.
Goeppert Mayer Fellow at Argonne National Laboratory. Connect transportation.gov/sites/dot.gov/files/2024-12/Quantum%20
with him on LinkedIn at https://www.linkedin.com/in/rshaydu. Workshop_Report_121024.pdf.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 23

WE NEED TO DEVELOP A CULTURE OF Digital Workers: AI at Your Service!
collaborative intelligence. There is no doubt that computers are increasingly
Responsible organizations are figuring out how capable of doing things that only humans could
to enable their top talent with smart computers, do in the past. Figure 1 shows the projected cost of
instead of replacing them. Leading organizations digital workers if Moore’s Law continues, in which
invest in empowering (augmenting) their employees an exascale is the estimated computing power of
with artificial intelligence (AI) so they can create the human brain. Advances in deep learning allow
higher value in better ways, as well as automating the digital workers of today to recognize speech
processes that are lower value and do not and images as well as process natural language to
engage employees. translate text and answer questions, not to mention Over time,
Every business is a talent business, right [1]? To advances in self-driving cars, emergency-response digital workers
stay ahead of the competition, every organization robots and robot assistants in stores. As machines
wants top talent who can learn faster, stay productive become smarter, talent management becomes can progress
longer, use technology more efficiently and invent more challenging for organizations because of the
higher-value offerings more effectively – we call this ever-growing range of knowledge-intensive tasks
from tool to
upskilling. Boosting the creativity, productivity and that can be automated and work processes that can assistant to
engagement of employees is a never-ending quest be transformed or newly created. Although digital
of innovation leaders in organizations as well as the workers can make a wider range of work processes
collaborator
talent management industry. more efficient, they can also create new challenges to coach to
Nevertheless, too many of today’s employees for organizations that want engaged employees to
are distracted, overwhelmed and impatient. In grow revenue and governments that want upskilled mediator,
the most competitive industries and geographies, citizens to grow GDP. augmenting
employee annual turnover rate for knowledge As Alfred North Whitehead said, “Civilization
workers can be more than 50% [2]. Millennial progresses based on the number of actions people both our own
and Gen Z “job-hopper” turnover costs the U.S.
economy $30.5 billion annually [3].
can do well without thinking about them” [4]. This capabilities and
may be true, but managing the transformation
Collaborative intelligence (people, AI and should and can be done in a way that upskills interactions
optimized processes) has the potential to improve engaged employees, rather than creating fear,
the efficiency, effectiveness, sustainability and uncertainty and doubt. From smartphone
with other
innovativeness of product and service offerings, assistants to driverless cars, the application people,
as well as disrupt business models, and can areas for applying artificial intelligence are
create new challenges and reasons for high- rapidly expanding, touching an ever-widening organizations
performing employees to learn and stay with range of areas of our personal and professional and systems.
their employers longer. lives. For example, OpenAI’s ChatGPT, Anthropic
In this article, we will first discuss the challenges Claude, Google Gemini, Microsoft CoPilot and
and opportunities surrounding the management other generative AI systems are reaching a level
of engaged knowledge workers as AI capabilities of utility that already provides a foundation
increase, and then we will provide guidelines to for a new generation of digital or cybernetic
develop a culture of collaborative intelligence teammates [5]. Over time, digital workers can
by more appropriately enabling automation progress from tool to assistant to collaborator
and augmentation. to coach to mediator, augmenting both our

$1T FIGURE 1: Petascale

Pe Ex supercomputers exist
ta as today, and by 2025, exascale
sc ca
al le supercomputers will exist with
e (1
(1 0*
roughly the same computing
$1B 0* *1 power as the human brain. If
Te *1
ra 5) 8) Moore’s Law continues as it has
sc in the past, by 2055, the cost of
al Super Computer Cost
e digital workers will approach
(1 $1,000 for narrow AI-enabled
0*
C

$1M Mainframe Cost

os

*1 petascale digital workers and,

t

2) 20 years later, broad AI-enabled

of

Person Average Annual

M G
D

exascale digital workers.

Ki eg
ig

ig Salary (Living Income)

it

lo as as
al

sc ca ca
W

$1K al le le
or

e (1
Smartphone Cost
ke

(1 (1
0*
rs

0* 0*
*3 *6 *9
) ) )
$1
1960 1980 2000 2020 2040 2060 2080
+/- 10 years

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 25

TALENT MANAGEMENT

SMART IMPROVEMENT TASK MODEL/ SELF MODEL/ USER MODEL/ WORLD MODEL/
MACHINES MODEL/ PERCEPTION/ REASONING/ INTERACTION/ KNOWLEDGE/
LEARNING/DATA VARIETY COMMON SENSE EPISODIC LEGAL TRUST
MEMORY
Tool + ++ + + +
Assistant ++ +++ +++ ++ ++
Collaborator +++ +++ +++ ++++ +++
FIGURE 2: Digital workers can
progress from tool to assistant Coach ++++ ++++ ++++ +++++ +++++
to collaborator to coach
Mediator +++++ +++++ +++++ +++++ +++++++
to mediator.

own capabilities and interactions with other Based on the current rate of progress, digital
people, organizations and systems. Advancing workers will be used by individuals on their
capabilities will allow digital workers to become personal upskilling journeys as (l)earners on more
trusted partners taking on routine, tedious tasks and more diverse platforms. In these scenarios,
and interactions on our behalf. This will lead individual employees will need to co-work with
responsible organizations to race to find durable the distributed intelligence of people, machines
competitive advantage that engages and upskills and organizations. Disruptive changes to business
employees [6]. Conversational AI is a great models that responsibly monetize employee and
example of these interactions using voice [7]. customer experience data will have a profound
Each of us will be on a personal upskilling journey impact on the employment landscape over
with a growing number of digital workers and their the coming years. Given these dynamics, how
sophisticated models operating on our behalf. In should talent management be transformed to
fact, by 2035, digital workers are expected to possess create a culture of collaboration intelligence
humanlike episodic memory systems – something that appropriately balances augmentation
no such system possesses today. Indeed, many and automation?
engineers predict that by 2055, each person may
have 100 digital workers that “work for them” [8]. Guidelines: Toward a Culture
While this may seem strange at first, consider of Collaborative Intelligence
farmers throughout history increasing the number The starting point is to understand task automation
of biological workers (plants and animals) working via artificial intelligence (smarter machines –
for them. Upskilled employees will be learning to digital workers) as well as interaction augmentation
responsibly manage more and more digital workers via cognitive computing (smarter people and
that operate on their behalf. organizations – upskillers growing their eminence).
As Moore’s Law continues and the number Don’t forget that cognitive computers do not
of digital workers per person increases, our “automate jobs” – they “automate tasks”! Here
expectations will increase, including boosting are some guidelines for creating a culture of
measures of interaction, creativity and productivity collaborative intelligence.
for both customers (users) and providers (employees). • An organization’s primary goal should
In fact, because everyone using digital workers will not be eliminating large numbers of jobs
be on a personal upskilling journey, responsible with cognitive technology. People and
firms operating platforms will see less distinction machines will work best in a relationship of
between customers and employees. Both will have augmentation (human in the loop) rather than
opportunities to earn more using the platform as full automation. AI addresses problems that
they upskill to better harness the power of digital can be neatly handled by digital workers –
workers. Both customers and employees are problems that are or can become reasonably
co-creators of value on the firm’s platform, and well defined and ultimately narrow in scope.
platform-enabled firms that take responsibility People excel at defining new problems that
for upskilling customers and employees will see need to be solved and coordinating with
a merging of marketing/communications (aimed other people to solve complex problems
at upskilling customers) and human resources with social dimensions.
functions (aimed at upskilling employees). • Substitution of digital workers for human
Almost all knowledge-intensive occupations, labor will be at the task level rather than the
including doctors, lawyers, investment bankers and job level for jobs that have complex social
policymakers, will need more digital workers built dimensions. For socially complex jobs, we have
from increasing capable and trusted cognition-as-a- been seeing more out-tasking of individual
service building blocks – personal tools, assistants, activities than outsourcing processes. Many
collaborators, coaches and mediators – that help jobs include structured, codified, routine and
augment and scale expertise (see Figure 2). predictable tasks that can be effectively

26 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 performed by computers, as well as tasks are needed for ubiquitous access and provision
in which people excel that involve emotions, of information among potentially unbounded
creativity, judgment, trust, empathy, ethics and networks of upskillers and their digital workers.
human intuition [9]. Finding the right balance Unstructured collaboration promotes creativity,
of cognitive computing and human skills while structured collaboration promotes operational
is important. efficiency. Most work will be done in project form
with adaptable workforces with an increased
Identify Non-value-adding Activities number of formal contractors and crowd platforms.
• Organizations need to be able to systematically It is extremely important to remember that the
scan entire portfolios of activities, workflows highest-performing employees look for new
and processes and categorize them according to challenges to intrigue their intellectual capabilities.
the business purpose, system interdependencies Smart technologies will learn how these people
and level of manual intervention required based work and can create new challenges for their
on value. They should be able to assess which learning and further development. There are always
tasks can be automated to reduce bottlenecks newer and better ways to address problems, and
and cost and also increase decision quality solutions to a problem can create new problems.
and compliance.
• Organizations must focus on enabling Develop a Culture of Lifelong Learning
employees – individuals on their upskilling Upskillers need to become analytical thinkers and
journey – to accomplish more value co-creating adaptive innovators. Research shows that digital
interactions with digital workers and encourage technologies change rapidly, but organizations
constant learning, driving relentless change. This and skills aren’t keeping pace. As a result, millions
will require them to adopt agile methodologies, of people are being left behind. Individuals need
service design, system thinking and open-source to familiarize themselves and learn the latest
community skills to enable employees to become capabilities of digital workers so they can become
adaptive innovators. close collaborators and value co-creators. For a
• Upskillers (talent with digital-age skills who range of tasks, children can still outperform digital
are constantly upskilling) are in great demand. workers, but for a growing number of tasks, digital
Winning and keeping them requires a concerted workers are faster, are more accurate and have
effort to build the values, beliefs, skills and a lower cost than hiring employees. People have
behaviors to support them. The change in the edge in commonsense reasoning tasks and
approach must extend to hiring practices and social intelligence, but digital workers are rapidly
accountability standards as well. New success improving and upskillers are exploring what
measures must be developed to reflect the is possible.
values of collaborating with digital workers. Training and education must be the core
Organizations need to signal a culture competency of organizations. The new breed of
of experimentation with the license to upskillers need to be lifelong learners in responsible
make mistakes. organizations that empower them to use the
platforms of their choice. Upskillers should also
Embrace Value Co-creating Interactions be encouraged to join professional development
Upskillers are sought after because of problem associations such as the International Society of
complexity and the pace of change. Complexity Service Innovation Professionals (ISSIP.org), Project
means the number of areas of knowledge that Management Institute, IEEE, INFORMS, ACM and
must be combined to solve problems is growing. many others. We learn best when we share our
Almost every unsolved problem today requires knowledge with others, which is one reason why
engineers, managers, behavioral and social scientists, open-source projects attract upskillers, and more
communications experts, and policymakers to work and more organizations encourage their employees
together. For example, consider driverless cars and to work on open-source projects like the Linux
what just one incident of a malfunction can create in Foundation AI and Data.
terms of a problem to be solved. The pace of change We learn best when we teach others.
is driven largely by rapid technological changes, but Organizations need to encourage their employees
demographic, social, economic, environmental and to contribute to others’ learning via mentoring
regulation changes also contribute. For example, or participating in programs that include other
storm surges along the coastline and flows of employees, students, faculty and industry mentors
migrants that result from environmental change who are working on grand challenges as part
can dramatically impact regions and the work of their lifelong professional development and
that people do from one year to the next. intellectual stimulation.
Today’s upskillers are evidence-based complex The move toward automation and augmentation
problem solvers with social-emotional intelligence with digital workers is a journey, requiring changes
and cognitive flexibility. Agility, collaboration and in both mindsets and behaviors. Succeeding in
skill sharing are also needed. Intelligent spaces such a journey requires building the community

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 27

TALENT MANAGEMENT

to support it. Individuals who are upskilling have IBM’s Venture Capital Relations Group in Silicon Valley. Jim has a Ph.D.
the abilities and interests to make organizations in computer science/artificial intelligence from Yale University and a
successful in this journey. They are broad, empathetic B.S. in physics from MIT.
communicators and challenge seekers with intuition,
REFERENCES AND NOTES
as well as deeply engaged critical thinkers with
1. https://time.com/charter/6247615/why-managing-talent-
collaboration and teamwork skills. They are is-now-finally-a-core-business-function/
entrepreneurial with a growth mindset, opportunity 2. https://www.bls.gov/jlt/home.htm
finders with imagination, who learn quickly from 3. https://www.businessnewsdaily.com/7012-millennial-job-
failure. We call them “T- or Pi-shaped analytical hopping.html
4. https://en.wikiquote.org/wiki/Alfred_North_Whitehead
thinkers and adaptive innovators.” It is also 5. Dell'Acqua, F., Ayoubi, C., Lifshitz, H., Sadun, R., Mollick, E., Mollick, L.,
extremely important to define the right et al., 2025, “The Cybernetic Teammate: A Field Experiment on
processes, metrics and measures, and to Generative AI Reshaping Teamwork and Expertise,” Working Paper,
capture, calculate, monitor and incentivize National Bureau of Economic Research, https://www.nber.org/
incremental productivity, efficiency and system/files/working_papers/w33641/w33641.pdf.
6. Bradford C. Johnson, James M. Manyika and Lareina A. Yee,
effectiveness of individuals, teams and 2005, “The next revolution in interactions,” McKinsey Quarterly,
organizations to enable adoption of AI for November 1, https://www.mckinsey.com/capabilities/people-
augmentation, automation and speed. and-organizational-performance/our-insights/the-next-
revolution-in-interactions.
HALUK DEMIRKAN is a Milgard Endowed Professor of Service Innovation 7. Ankush Sabharwal, 2023, “The next revolution in tech: What to
& Business Analytics, University of Washington Tacoma, and a head of know before implementing conversational AI,” Forbes, February 27,
cognitive science at Amazon Devices & Services Organization. He is the https://www.forbes.com/councils/forbestechcouncil/2023/02/27/
co-founder and board director of the International Society of Service the-next-revolution-in-tech-what-to-know-before-implementing-
Innovation Professionals. He is a Distinguished Applied Research conversational-ai/.
Scientist with nearly 25 years of expertise in advanced analytics, 8. Jim Spohrer, Md Abul Kalam Siddike and Youji Kohda, 2017,
machine learning and artificial intelligence. Haluk has a Ph.D. “Rebuilding evolution: A service science perspective,” Proceedings
in information systems and supply chain management from of the 50th Hawaii International Conference on System Sciences,
University of Florida. https://scholarspace.manoa.hawaii.edu/bitstream/10125/41354/1/
paper0205.pdf.
JIM SPOHRER is a board director of ISSIP & ServCollab and retired 9. Zhou, L., Paul, S., Demirkan, H., Yuan, L., Spohrer, J., Zhou, M. &
director, Cognitive Opentech Group, IBM Research – Almaden, San Jose, Basu, J., 2021, “Intelligence augmentation: Towards building
California. Previously, Jim helped to found IBM’s first Service Research human-machine symbiotic relationship,” AIS Transactions on
group, the global Service Science community, and was founding CTO of Human-Computer Interaction, Vol. 13, No. 2, pp. 243-264.

INFORMS’ LATEST
PUBLISHED BOOK
AVAILABLE NOW!
Dick Larson’s accessible and
innovative approach to model
thinking. Scan QR code to purchase
your copy from Amazon!
 TAP THE SMARTEST MINDS IN
O.R., ANALYTICS & DATA SCIENCE –
OCTOBER 26–28 IN ATLANTA

Whether you’re building your analytics team or expanding your

academic department, the INFORMS Career Fair gives you early
access to the brightest minds in operations research, AI, and data
science. Held October 26–28 during the INFORMS Annual Meeting,
this is your opportunity to meet top talent, access resumes, and build
your pipeline – before your competition does.

WHY PARTICIPATE?
The Career Fair is an ideal place to find new team members. Whether
you’re a national or international organization or seeking academic
or industry talent, this event can help you source the talent you need
to succeed.

At this event you can:

• Raise the visibility of your organization.
• Provide recruitment materials and display your brand.
• Fill your talent pipeline and open positions as you meet
and interview qualified candidates.
• Tap into the INFORMS Career Center resume database.
• Use the INFORMS Career Fair scheduling tool.

Secure Your Spot Before Booths

Sell Out – Register Today
OPERATIONS RESEARCH (O.R.) WAS BORN The Role of AI in O.R.: Existential Threat
during World War II, when interdisciplinary or Useful Tool?
teams of mathematicians, scientists and military With the latest rise in AI techniques and
personnel worked together to solve complex applications, the impact of AI on society and
logistical and strategic problems. Since then, other scientific areas is fiercely debated. Recently,
O.R. has helped society make smarter decisions, Ramayya Krishan (Carnegie Mellon University
optimizing decision-making in industries such as and 2016 UPS Smith Prize winner) summarized
manufacturing, healthcare, telecommunications this debate in two schools of thought: AI is seen
and transportation. For decades, O.R. graduates as just another technology that is adding tools in
have specialized in mathematically modeling the toolbox; the other direction points to a more
real-life systems by delving into their operations, significant disruption as we realize artificial general
analyzing the processes involved in describing and intelligence. The panel discussion reflected this
predicting their performance, optimizing processes divide, underscoring the challenges of reaching a
and the design of systems, and designing control consensus on AI’s future impact.
mechanisms to maintain the desired performance Those viewing AI as a tool emphasize that O.R.
levels and adapt to changes or uncertainties. is a methodological approach to system analysis,
A decade after the birth of O.R., the work of optimization and control, which borrows and
Alan Turing on symbolic reasoning systems is often develops various tools to achieve this goal. Vlasiou
taken as the birth of artificial intelligence (AI). In pointed out that simulation was seen in a similar
its early years, the limited computational power light in the past, but the outcome has been that How will AI
hindered progress, with the main developments simulation techniques pushed the boundaries
focusing on symbolic AI. Overhyped promises failed of O.R. and helped sharpen its impact. In her impact O.R.
to deliver, leading to the first AI Winter in the 1970s. view, “although AI is seen by some as a panacea
After an AI revival and subsequent period when to solve all problems, it is evident that it is a
practitioners?
the promise of theory struggled to meet practical tool that performs better than existing tools in Is it simply
needs and financial support dried up, computer some respects, such as recognizing patterns and
scientists, such as Geoffrey Hinton, Yann LeCun generating predictions based on large complex another tool
and Yoshua Bengio, persevered in their work on datasets, but is challenged in other areas, such that can be
neural networks and deep learning. The growth in as creative design or navigating nuanced human
computing power and data availability has brought emotions. I am sure we will keep pushing these added to an
breakthroughs in image and speech recognition, boundaries, too.” extensive and
natural language processing, large language models Avoiding this dual divide, Krishnan offered a
and more. Currently, AI is widely adopted across synergistic outlook. C. West Churchman – the proven toolbox,
industries and is broadly viewed as one of the most founding editor of Management Science – defines
transformative technologies of our time and the O.R. as “securing improvement in social systems by
or is it poised to
next driver of innovation that will shape our future the scientific method” [1]. Rather than viewing AI as fundamentally
in unprecedented and unpredictable ways. a threat, Krishnan argued that “the job of creating
How will AI impact O.R. practitioners? Is it simply that mutual understanding between analysts and reshape our
another tool that can be added to an extensive and decision-makers – and therefore the work required approach
proven toolbox, or is it poised to fundamentally for successful implementation of decisions and
reshape our approach to solving problems? In either the improvement of social systems – still requires to solving
case, how should O.R. educators and practitioners human skills.” In this context, O.R. practitioners problems?
respond to AI and its impact? should, therefore, be particularly well placed to
A panel of experts, comprising past winners future-proof themselves by identifying the areas
of the INFORMS UPS George D. Smith Prize and in which human input is required.
moderated by Anne Robinson and Ranganath Although the well-established concept of
Nuggehalli, debated these questions during the “human in the loop” remains a crucial factor in AI
2024 INFORMS Annual Meeting in Seattle. The UPS systems today, the future of AI may evolve beyond
Smith Prize recognizes an academic department this dependency. Zhou contended that although AI
or program for effective and innovative preparation is not an existential threat to O.R. practitioners yet,
of students to be outstanding practitioners of it is “developing at an unprecedented rate. Current
operations research, management science, analytics generative AI tools are not the best at crunching
and data science. The panelists – Michael Fry numbers or logical reasoning. However, given
(University of Cincinnati), Harish Krishnan how fast these technologies are developing, it
(University of British Columbia), Maria Vlasiou may further disrupt every field, including O.R.”
(University of Twente), Bradley Warner (United At the other end of the spectrum are those who
States Air Force Academy) and Wenjun Zhou view AI as a general-purpose technology with the
(University of Tennessee, Knoxville) – discussed the potential to revolutionize all aspects of our lives,
role of AI in the context of O.R., how it can shape O.R. including medicine, law, retail, entertainment and
education and how to future-proof O.R. graduates. communication. According to Warner, “Generative

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 31

O.R. PRACTITIONERS

AI is neither simply a tool nor an existential threat, unforeseen implications for an administration
but a profound disruptor that will transform not not used to perfect outcomes throughout the
only analytical processes but the entire landscape whole cohort.
of decision-making.” The University of Cincinnati launched an
According to Vlasiou, “the discussion of whether Applied AI Lab to serve as a comprehensive
a new technology is an existential threat to the resource hub for generative AI, educating faculty
status quo is an old one,” drawing analogies to and staff across departments on generative AI
the old debate that the use of pocket calculators capabilities to incorporate AI into courses in
and computers would hamper the development meaningful ways and actively supporting cross-
of computational skills. The consensus was that disciplinary integration for courses across majors.
O.R. education programs need to innovate the In addition, the lab has partnered with companies
curriculum by integrating AI into their programs, not only to help upskill their workforce but also
exploring its limits and navigating its challenges. to collaborate on applying generative AI to solve
real-world, practical problems – offering valuable
Integrating AI into Education learning opportunities for both students and faculty.
Although AI is a field that has existed for more than Whether cautious or wholehearted, the integration
70 years, its widespread impact on education is of AI into education comes with the need to carefully
only now emerging. AI is impacting not only how evaluate policies and procedures. As Vlasiou notes,
a subject is taught but also what is taught. The “no matter what policy regulations exist, students
different viewpoints on the role of AI – whether it will ask AI for help with their tasks.” This spurred
should be viewed as yet another useful tool for the the Eindhoven University of Technology to adapt
service of science, technology and society or as policy and regulations to reflect this new reality.
a major disruptor to every imaginable field – are Krishnan expects that many tasks currently
also reflected in its adoption across educational requiring human input will become increasingly
programs. The panel offered perspectives and automated and maintains that although “AI tools
examples of how AI reshapes the curriculum, should be integrated into educational programs,
delivery method and role of industry. they must be positioned as a complement to
The Business Analytics program at the University human skills rather than a replacement.”
of Tennessee, Knoxville, has been quick to update
its curriculum in response to market changes. The Path to Ethical AI
Before this new wave of generative AI, it already Ethics have always been paramount in data-driven
offered in-depth content on deep learning and decision-making, although its complexity has
reinforcement learning. The content of existing evolved significantly over time. Initially, ethical
courses is also adapted to market and student considerations were more straightforward –
needs. Using her text mining class as an example, focusing on understanding data sources, ensuring
Zhou introduced the use of AI tools such as data cleanliness, assessing generalizability and
Open AI’s API and noted the advantages and being aware of tool-specific assumptions.
disadvantages of using AI to perform text mining As technology progressed, ethical concerns
tasks that were previously done with coding. As AI expanded. With the advent of computers, we
impacts the future of work, it also shifts what we needed to address data privacy, security and the
teach to prepare students. appropriate use of computational algorithms.
The Data Science program at the United States With analytics and AI, newer ethical concerns
Air Force Academy views AI as a collaborative include transparency, fairness and accountability.
partner that augments the capabilities of lecturers For example, AI algorithms can amplify biases and
and enhances efficiency. Rather than being a affect groups in profound ways. Models with high
“simple” automation tool, AI is used as a teaching predictive accuracy may not be good models if they
assistant, providing detailed and personalized make unfair decisions because they were trained
feedback that is straightforward to produce and with biased data. Black-box models with high
valuable to the learner, as validated by human complexity may not be trustworthy for high-
instructors following the human-in-the-loop stake applications.
principle. Education was transformed into a truly Generative AI introduces a nuanced ethical
outcome-based learning experience, with students situation that encompasses all previous concerns
gaining valuable experience that enhanced their while introducing novel challenges. According to
own learning process. Furthermore, the role of Warner, what is new to generative AI is “the subtle
lecturers shifted from adjudicators to mentors, and potentially dangerous tendency we may have
shaping the interactions between lecturers and to over-trust AI systems because they appear to be
students. Rather than assessing students at human-like and they are also authoritative.” Zhou
specific moments, continuous feedback changed stressed that educators “share the responsibility of
the approach of students to education: Students informing students of generative AI’s limitations,
persisted until they succeeded, leading to such as hallucinations and inadequate reliability,

32 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

which might make it unfit for high-stake part of O.R. practitioners’ work. However, the
applications.” Fry added, “Our students must implementation part still requires human input.
learn how to detect, assess and mitigate bias in We should recognize this distinction and that we
AI – a skill that should be as essential as measuring face a choice in how we incorporate AI tools into
model accuracy. Teaching students to examine our systems. While there is a risk of misusing AI
AI’s societal impacts is key to responsible decision- by excessively relying on automation instead of
making in practice.” leveraging its potential in proper use cases as a
The ethical imperative remains clear: We must tool to augment human labor, learning how to
always keep humans in the loop, recognizing that use it correctly and responsibly seems to offer
certain skills and attributes are fundamentally and greater benefit.
distinctively human. As AI takes on routine tasks, Future-proofing O.R. practitioners demands
human decision-making, especially in ethically more than adapting to tools – it requires strategic
complex situations, becomes more valuable. shifts in what we teach, how we learn and how we
O.R. professionals must be equipped to make work. For academics, this means going beyond
nuanced judgments about how AI is deployed introducing AI as a topic in a single elective or
and ensure that it upholds ethical standards. It is the module. It involves thoughtfully embedding AI
responsibility of O.R. professionals to monitor the literacy, ethics and hands-on experimentation
outcomes of AI-driven decisions, particularly when across the curriculum while maintaining a strong
these decisions have far-reaching effects on people, foundation in core O.R. principles. Faculty should
communities or the environment. design assignments that require students to
both apply and critique AI-generated solutions
Future-proofing O.R. Practitioners and explore partnerships with industry to create
As education shifts with the dominance of and experiential learning opportunities rooted in
easy access to AI, changes are already visible in the real-world AI challenges. Just as importantly,
workplace. Traditional programming skills could academic programs must prepare students
become obsolete as AI may offer a humanlike way to navigate the organizational, social and
to communicate with computers. With direct ethical dimensions of deploying AI-enabled
human-computer communication becoming more systems – skills essential for meaningful and
accessible, it is unclear whether fundamental responsible implementation.
programming knowledge remains essential. For students, it involves mastering both
The human-in-the-loop principle indicates that the fundamentals of O.R. and the evolving
fundamental skills continue to be important. capabilities of AI while developing the ability to
Arguably, generative AI tools will “flatten” the operationalize solutions – translating analytical
difference between highly skilled and less skilled insights into actionable, real-world impact.
programmers; as seen throughout history, other This means going beyond model building
disruptive technologies have not eliminated the to understand implementation constraints,
need for fundamental understanding. Elementary organizational dynamics and the iterative
students still need to learn basic arithmetic, process of embedding solutions into practice,
despite the wide availability of calculators and where human judgment and contextual
computing software. awareness remain essential.
Regardless of the precise balance, human For practitioners, future-proofing means not
workers will coexist with AI for some time. In only keeping up with advances in AI but also
broad lines, skills such as domain knowledge, working effectively across disciplines – bringing
interdisciplinary understanding, problem-solving, together operations researchers, data scientists,
critical thinking, modeling, mathematical analysis software engineers and domain experts to
and computational skills will remain critical. Even develop solutions that are technically sound,
if AI is able to perform some analytical tasks “better” implementable and aligned with organizational
or faster, O.R. graduates should be in a position goals. These interdisciplinary teams are essential
to check this outcome, validate it, explain it and because deploying AI in real-world settings
understand how it came to be. The premium for requires more than just a good model: It demands
human workers will remain in those tasks (e.g., thoughtful integration into workflows, attention
strategic thinking and coordination), where AI is to user needs and ongoing evaluation – areas
more likely to augment human decision-makers where collaboration across expertise is crucial.
rather than replace them. Across all roles, the goal remains the same: to
The adaptation to a world filled with AI is not a use AI not as a replacement for human judgment
binary question. Early scholars and practitioners but as a catalyst for amplifying it.
of O.R., including C. West Churchman and others,
were extremely prescient for noting the distinction REFERENCE
between analysis and implementation. AI tools 1. Churchman, C.W., 1970, “Operations Research as a Profession,”
will certainly enhance and speed up the “analysis” Management Science, Vol. 17, No. 2.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 33

O.R. PRACTITIONERS

MICHAEL FRY is a professor of operations, business analytics and

information systems at the University of Cincinnati as well as director
of the Lindner Centers & Institutes.

HARISH KRISHNAN is Advisory Board Professor of Operations and

Logistics at the Sauder School of Business, University of British Columbia.
Harish received his Ph.D. in business administration from the University
of Michigan.

MARIA VLASIOU is a full professor of stochastic processing interacting

networks in the Faculty of Electrical Engineering, Mathematics and
Computer Science at the University of Twente (UT), director of the Dutch
Network on the Mathematics of Operations Research (LNMB) and a
Research Fellow of the European research institute EURANDOM.

BRADLEY WARNER is a professor of applied statistics at the United States

Air Force Academy. His research interests include machine learning, neural
networks and mathematical pedagogy.

WENJUN ZHOU is the Lawson Professor of Business, Ph.D. in Business

Analytics Program Director, and Martin Lee and Carol Fri Robinson Faculty
Fellow at the Haslam College of Business, University of Tennessee,
Knoxville. She currently serves as the president of INFORMS College
on Artificial Intelligence.

ANNE G. ROBINSON is a renowned supply chain strategy and analytics

executive with expertise in digital transformation. She previously served
as chief strategy officer at Kinaxis, and led global supply chain strategy and
analytics at Verizon and Cisco. She is an INFORMS Fellow and past president.

RANGANATH NUGGEHALLI is a retired principal scientist for UPS. He is an

INFORMS Fellow and was part of the UPS team that won the 2016 Franz
Edelman Award for UPS’ On-Road Integrated Optimization and Navigation
(ORION) Project.

Check out INFORMS Analytics magazine,

the leading online publication for
analytics professionals.

www.analytics-magazine.org
FIND YOUR PEOPLE. FUEL YOUR POTENTIAL.
INFORMS Forums for Early-Career Professionals
Looking to connect, collaborate, and grow within the OR/MS community? INFORMS offers
seven vibrant forums designed to unite members around shared professional interests beyond
discipline or geography.

If you’re an early-career professional, JFIG and JRING are your launchpads:

• JFIG – Junior Faculty Interest Group
Connect with fellow junior faculty, senior mentors, and industry leaders. Build your
network. Spark new research collaborations. Navigate your academic journey with a
strong support system.

• JRING – Junior Researchers in Industry, Nonprofit, and Government

Join a community that thrives on real-world impact. JRING brings together early-
career researchers working outside academia to share insights, challenges, and
opportunities in OR/MS.

Professors, spread the word! Let your students and junior colleagues know about these
valuable communities – they’re a game-changer for career development and connection.

Find your forum. Join today:

https://www.informs.org/Communities/INFORMS-Forums

Also check out:

The CTL Teaching Excellence and Networking (CTL-TEN) workshop is an evolution of the
Early Career Teachers’ Network (ECTN) initiative, designed to serve a broader community
of academics passionate about teaching OR/MS and analytics. This preconference event at
the 2025 INFORMS Annual Meeting offers opportunities to engage as a mentor, mentee,
participant, or panelist/speaker.

Join us on October 25, 2025, for a day of mentorship, professional development, and
meaningful conversations about teaching.

https://meetings.informs.org/wordpress/annual/ctl-ten/
SINCE THE PUBLICATION OF THE BOOK Strikeout % (K%) and Walk % (BB%) Over Time
“Moneyball” in 2002 (and subsequently, the 30
2011 movie adaptation), baseball has shifted
from an era of intuition – relying on gut feelings
and traditional knowledge – to a scientific era 23.2
22.4 22.7 22.6
defined by data-driven decision-making. By 2021,
analytics had evolved beyond basic statistics into
20
multidimensional systems powered by advanced

Percentage
technology, as detailed in the article “From Metric
Moneyball to Betterball” [1]. K%
This transformation accelerated further with BB%
innovations like the pitch clock, Automated
Ball-Strike System (ABS) and biomechanical 10 8.7
8.2 8.6 8.2
tracking platforms. This article explores two key
questions: What groundbreaking innovations
have revolutionized baseball analytics since 2021?
How have these changes reshaped the essence of
the game? 0
2021 2022 2023 2024
Recent Technological Innovations and Season
Rule Changes
In 2023, the ABS underwent two significant players such as Corey Seager and Shohei Ohtani FIGURE 1: Change in strikeout
modifications at the Triple-A level. The first three experienced the most substantial gains, with their percentage (K%) and walk
percentage (BB%) from 2021
games of a series implemented a fully automated OPS (on-base plus slugging) increasing by as much
to 2024.
strike zone to determine balls and strikes, whereas as 271 points and their weighted on-base average
the subsequent three games employed a “challenge” (wOBA) rising by over 100 points.
system. In this format, the umpire made the initial However, teams quickly devised countermeasures.
call. However, the hitter, picther or catcher could Some elite teams adjusted outfield positioning,
challenge the decision up to three times, retaining adopting a “shallow right field + deep left field”
each challenge if correct, with the ABS providing hybrid alignment to partially offset the effects of
the final verification. This challenge system was the rule change. The Houston Astros, for example,
piloted in select stadiums during the 2022 season leveraged StatCast-generated heat maps of batted-
and was well received, prompting Major League ball distribution to optimize defensive alignments.
Baseball (MLB) to expand its use in 2023 to In the 2023 season, they successfully reduced right-
determine the optimal format preferred by handed hitters’ opposite-field hit rate to a league-
players and coaches. low 9%, demonstrating how defensive strategies
High-speed cameras and 3D video tracking have rapidly adapt in the era of data-driven baseball.
revolutionized player biomechanics by providing Meanwhile, MLB’s introduction of an automated
real-time motion analysis [3]. Teams now optimize challenge system has also profoundly altered the
pitching mechanics, limb loading and pitch trajectory; competitive landscape. This system, which relies on
monitor recovery progress; and mitigate injury the Hawk-Eye optical tracking system for real-time
risks with unprecedented precision, giving them officiating, has significantly improved the accuracy
a competitive edge. For example, Donna Moxley of umpiring decisions. Although these two reforms
Scarborough, a biomechanics expert for the Boston target different aspects of the game, both exemplify
Red Sox, has leveraged 3D video technology to how data-driven decision-making is reshaping
evaluate prospective pitchers, refine pitching modern baseball.
mechanics and design personalized rehabilitation Shift restrictions have altered the balance between
programs for players [4]. offense and defense through rule adjustments, and
The 15-second pitch clock reduced average game the pitch clock has urged pitchers to make quicker,
length by over 30 minutes in 2023, enhancing the more effective decisions. Notably, the ability for
pace of play without significantly changing the walk teams to quickly adapt – such as the Houston Astros’
and strikeout rates, as shown in Figure 1. Pitchers innovative defensive adjustments – reinforces
were able to maintain effective control, even with a crucial insight: The evaluation of any policy’s FIGURE 2: Left-handed batter
performance in 2022 and 2023
reduced preparation time. effectiveness must consider not only the direct shows the impact of the 2022
Shift restrictions in 2022 restored traditional rule change but also the behavioral adaptations shift restrictions on offensive
defensive alignments by prohibiting infielders and tactical evolutions it triggers. From a broader performance metrics.
from clustering entirely on one side of the field [5].
Data pulled from FanGraphs “Splits Leaderboard” Season BABIP AVG OBP SLG wOBA
shows that this policy significantly improved
2023 0.297 0.249 0.328 0.418 0.323
offensive performance for left-handed hitters [6],
2022 0.287 0.239 0.315 0.391 0.310
as presented in Figure 2. Notably, high-shifted

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 37

BASEBALL ANALYTICS

perspective, these reforms are not merely altering formal integration of collegiate training data into
the mechanics of individual games; they are professional scouting processes. Using advanced
driving baseball toward greater fairness flight tracking systems, programs at the University of
and precision. Florida and Wake Forest University provide hitters
and pitchers with real-time feedback on metrics such
Emerging Data Sources: as exit velocity (accuracy: ±0.3 mph), launch angle
From Laboratory to the Field (resolution: 0.1°), spin direction, seam displacement
Fan Interaction Analysis [7] (horizontal/vertical deviation error: <0.5 cm) and
spin axis deviation angle.
Democratization Technological innovation is transforming fan
experiences in baseball, making them more This data-driven approach has transformed
of analytics personalized and immersive. Teams use sentiment player evaluation and development. For example,
analysis of social media to gauge fan reactions former University of Tennessee slugger Blake
has leveled the to key game events, such as no-hitters or hitting Burke improved his slugging percentage by 0.081
playing field, for the cycle, enabling rapid content generation. by optimizing his launch angle based on Rapsodo’s
Machine learning models analyze data such as insights into the “golden launch zone” (25°-35°).
empowering ticket purchases and merchandise preferences Similarly, pitcher Brody Brecht’s anomalous slider
teams at to tailor fan engagement strategies. movement (7.2 cm lateral at 1,850 rpm) propelled
Augmented reality enhances live broadcasts by him from a second-round projection to first-
all levels to overlaying real-time data, such as pitch velocity round contention.
compete in and launch angle on top of the live game, catering Rapsodo’s affordable systems have democratized
access to advanced analytics for mid-tier programs
to analytically inclined fans. Interactive features
this new era of like prediction polls further engage audiences. while bridging gaps between professional and
Looking ahead, advancements in edge computing collegiate levels. This shift suggests MLB scouting
precision-driven and cognitive science – such as eye-tracking is evolving from an experience-driven model to
baseball. technology – promise to optimize in-stadium one powered by algorithms.
experiences by aligning content delivery with
fan attention patterns. Challenges and Controversies
AI Umpires vs. Human Umpires
Expansion of Collegiate Baseball Analytics [8] Umpire errors introduce randomness into crucial
Rapsodo’s sensor technology is revolutionizing talent moments, potentially undermining competitive
development in NCAA baseball. Its partnership fairness. Data pulled from Umpire Scorecards
with ESPN for the 2024 MLB draft marked the first (https://umpscorecards.com/data/teams), and

FIGURE 3: Scatterplot shows a MLB Team Winning % vs. Total Favorability (2024)
weak but positive correlation
between MLB teams’ 2024
winning percentages and 0.6
their total favorability from
incorrect umpire calls.

0.5
Winning Percentage

0.4

r = 0.383

0.3

-20 0 20
Total Favorability

38 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

portrayed in Figure 3, shows there is a weak but injury risk. For instance, a Yankees minor league
positive correlation (r = 0.383) between MLB pitcher improved his slider mechanics based
team winning percentages in 2024 and the net on AI recommendations, increasing horizontal
run impact caused by incorrect calls by umpires. movement while minimizing elbow strain.
Although this statistical relationship is not Deep learning models are transforming game
significant, the revealed trend remains noteworthy preparation by decoding opponents’ strategies.
because it suggests that even small advantages AI simulations analyze pitchers’ tendencies,
gained from incorrect umpiring decisions could batters’ weak zones and even umpires’ strike
possibly influence game outcomes. zone patterns to help coaching staff rehearse
Although not a decisive factor in most games, scenarios and develop targeted tactics. Teams
incorrect calls can subtly influence outcomes over gain a competitive edge by integrating these
a season. The ABS has reduced human umpire insights into pregame planning.
error rate from 14% to less than 2%, significantly AI is reshaping scouting by replacing subjective
improving accuracy and mitigating biases. evaluations with predictive models that integrate
However, concerns about the “human element” data on metrics including exit velocity, base-
remain. A hybrid approach, such as the limited running efficiency and defensive reaction time.
challenge system trialed in minor leagues, combines Generative AI tools simulate long-term player
technology with tradition. Teams are allowed performance trajectories, assisting teams in draft
three manual reviews per game, ensuring over 95% decisions and contract negotiations. This shift
accuracy while maintaining tactical opportunities toward data-driven scouting enhances accuracy
for players. A 2023 minor league trial of this system in identifying high-potential talent.
saw a 41% increase in live-game approval ratings.
Ultimately, artificial intelligence (AI) umpires Breaking Physical Boundaries for
aim to restore fairness by ensuring that game Athletes and Fans
outcomes are determined by player skills – with Virtual reality (VR) technology allows players to
bats and on the pitching mound – rather than train in realistic virtual environments, regardless of
subjective calls, striking a balance between weather or field availability. Batters can practice
technological precision and the emotional against simulated high-velocity virtual pitches,
appeal of human judgment. and fielders can rehearse catching scenarios under
varying stadium lighting conditions. For example,
Traditional Scouting vs. Data Analytics the Los Angeles Dodgers conducted a study
Traditional scouting methods, based on showing that VR training reduced rookies’ batting
subjective assessments like natural instincts reaction times, highlighting its potential to
or “mechanical potential,” are increasingly enhance player performance.
challenged by advanced analytics. Wearable Augmented reality (AR) is transforming how
tracking systems and computer vision now fans experience baseball. Spectators can use
quantify player attributes – such as release smart glasses or mobile apps to view real-time
point displacement and spin efficiency – visuals, such as exit velocity, pitch trajectories and
with precision, transforming vague talent defensive alignments, during gameplay. MLB’s
projections into measurable metrics. Edge-AR system overlays real-time biomechanical
The democratization of data is reshaping data – such as pitcher elbow valgus torque and
scouting roles. Affordable tracking systems now batter bat-head acceleration – onto live broadcasts,
allow mid-tier programs to uncover hidden providing an interactive and analytically rich
attributes such as seam-shifted wake effects, viewing experience.
forcing scouts to evolve from observers into
data interpreters. Machine learning models Injury Prevention and Load Management
highlight weak correlations (r < 0.35) between Subdermal nano-sensors now track microdamage
traditional scouting terms and performance in tendons, alerting teams when a pitcher’s rotator
indicators, pushing talent evaluation toward cuff strain exceeds a safe threshold. Additionally,
predictive science. smart protective gear can restrict throwing motions
to prevent further injury. These advancements mark
Baseball Ecosystem Transformation baseball’s evolution into a “super laboratory” where
AI and Machine Learning to Enhance computing power, data science and biotechnology
Coaching and Team Management converge to enhance player safety.
AI-driven training systems leverage wearable Dynamic load management integrates real-time
devices to collect real-time physiological data, biometric sensor data (e.g., heart rate variability,
such as heart rate, muscle load and motion muscle fatigue index) with external factors
trajectory. Machine learning algorithms analyze like weather and travel schedules to generate
this data to create personalized training programs individualized game participation strategies. For
that optimize performance while reducing example, if a pitcher experienced reduced sleep

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 39

BASEBALL ANALYTICS

because of cross-time-zone travel, the system promise to push boundaries even further. Yet,
adjusts their workload by altering pitch types or amidst these advancements, one thing remains
recovery protocols such as cold therapy. Similarly, constant: baseball’s ability to captivate players
in extreme heat or humidity, outfielders’ defensive and fans alike with its timeless appeal. As the
strategies are modified to conserve energy and game continues to evolve, its essence endures –
reduce injury risk. This adaptive model balances a celebration of skill, strategy and passion that
injury prevention and peak performance. no algorithm can replicate.

Technological Democratization Note. The reference list can be found online:

Low-budget teams can use open-source algorithms https://doi.org/10.1287/orms.2025.02.07.
to identify a prospect with a “low-spin, high-
movement” breaking ball. This fundamentally ZIYU (JACK) CHEN is an M.S. student in sport management at the
University of Florida (UF).
disrupts the traditional scouting model – once
reliant on expensive tracking systems. TRISTAN ROBBINS is an undergraduate student studying sport
MLB’s release of the Statcast Lite dataset management, statistics and economics at UF.
and open-source toolchains (Sabermetrics.py,
OpenPitch) is breaking resource barriers in WARNER GEPHARDT is an undergraduate student studying statistics
dynamic load management. Minor league and sport management at UF and leads the UF Sports Analytics Club.
teams collect pitcher joint torque data using SCOTT NESTLER, Ph.D., CAP-X, PStat, is professor of practice and sport
consumer-grade wearables. Data is then analyzed management at the University of Florida and director of the UF Sports
with federated learning models shared on GitHub, Analytics Lab.
enabling cost-effective load evaluation comparable
to elite sports labs. Even high school coaches can
analyze the correlation between launch angles and
weather conditions, designing seasonal-specific
training regimens – such as adjusting swing
balance for wet-field conditions – to stabilize
slugging percentage fluctuations.
Similarly, small-market teams can leverage a
shared Amazon Web Services cloud-based database
containing more than 200,000 hours of motion
analysis to obtain training insights comparable
to elite franchises at just 1/50th of the cost. This
access shifts dynamic load management from
an “elite privilege” to a universal tool, propelling
the entire industry toward an era of data-
driven accessibility.

The Future of Baseball in a Data-Driven Era

Baseball has always been a sport rooted in
tradition, but its evolution over the past two
decades underscores the transformative power
of data and technology. From automated strike
zones to dynamic load management systems,
these innovations are redefining how the game
is played, coached and experienced by fans. The
democratization of analytics has leveled the playing
field, empowering teams at all levels to compete in
this new era of precision-driven baseball.
However, as the sport becomes increasingly
reliant on technology, it must navigate challenges
such as preserving the human element and
balancing fairness with tradition. The integration
of AI umpires and algorithmic scouting reflects
this delicate balancing act – one that seeks to
enhance the game without losing its soul.
Baseball stands at the intersection of science
and sport, where breakthroughs in biomechanics,
machine learning and immersive technologies

40 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 REGISTER
TODAY!
2025 INFORMS
International Meeting

Singapore | July 20–23

https://meetings.informs.org/international

THANK YOU TO OUR SPONSORS!

Platinum Sponsor

Gold Sponsors

Silver Sponsor

Sponsors as of June 11, 2025. For the most up-to-date information, visit
https://meetings.informs.org/wordpress/2025international/current-sponsors/.
AS THE NUMBER AND INTENSITY OF 2. Hurricane Crisis Committee: Proactive
Atlantic hurricanes continue to increase, supply Planning and Flexibility
chain professionals must rise to the challenge and The existence of a company’s hurricane crisis
innovate in the face of a rapidly changing climate [1]. committee is essential for an efficient and swift
The logistics-related disruptions caused by response to any major supply chain disruption,
a hurricane can be broad-reaching, including including hurricanes. Hurricane crisis committee
limitations in transport, damage to warehousing members should have an in-depth understanding
infrastructure and downed communication lines, of each segment in the transport chain, from
as well as the potential strain on inventory port or factory to warehouse, transport and
availability. Transport limitations can result in customer. Understanding how these entities
delayed shipments, unpredictable transit times respond to disruptions, such as transportation
and reduced inventory levels at customer locations. interruptions, is crucial for creating contingency
Damage to warehousing infrastructure further plans to maintain service level requirements.
restricts the locations from which inventory can These plans should include identifying
be held and shipped, potentially causing additional obstacles and their respective solutions
transit time fluctuations. Additionally, downed to provide a quick reference for rapid
communication lines can create uncertainty and response. Additionally, the effectiveness
concern regarding the next steps, unless a seasoned of these committees is enhanced through
and nimble team is prepared to handle such events. regular scenario testing. Such practice runs
The timing of hurricanes can vary significantly, allow for developing creative solutions
with some storms forming rapidly and others giving and alternatives and strengthening team Accepting that
more advance warning. This variability necessitates rapport and innovation, even under high-
a flexible approach to planning and response. For pressure environments. hurricanes
example, Hurricane Milton in 2024 went from a 3. Use of Technology are natural
tropical depression to a Category 5 hurricane in Predictive analytics provides a wealth
49 hours [2], and the rapid formation and path of of information that should be regularly phenomena
the storm caused major disruptions throughout and readily involved in hurricane crisis
Florida [3]. This weather event required swift management efforts. Existing and regularly
predicted
and calculated actions to protect operations at monitored cloud-based systems, such as to occur
ports on the west coast of Florida and anticipated transport management systems (TMS) and
closures at terminals on the southeast coast. enterprise resource planning (ERP) systems,
with higher
The unpredictability and volatile nature of these should be creatively integrated with real-time frequency and
phenomena imply that logistics professionals weather prediction models to assist with
must always be on high alert, ready to implement rapid, prediction-based planning related to intensity is
contingency plans immediately. hurricanes. These tools improve the modeling a necessary
of contingency plans before hurricane events
Risk-Mitigation Strategies and better prepare alternative plans as the first step in
To effectively manage these disruptions, creating events unfold. By using these tools, potential hurricane
a hurricane crisis committee that can study and bottlenecks and obstructions in the supply
implement strategies to mitigate associated risks chain can be identified, and alternatives can supply chain
is pertinent. These proactive strategies encompass be rapidly suggested. For instance, companies
a comprehensive range of topics, from planning utilizing railcar tracking tools from Quality
preparedness.
to coordination with cargo receivers, and aim to Transportation Services (QTS) can obtain
reduce the impact of hurricanes on breakbulk specific in-transit route reports, enabling them
supply chains. The following strategies suggest to anticipate railcars that might be impacted
an approach to increasing resilience in the face by a hurricane. The hurricane crisis committee
of hurricanes. can then combine these suggestions with their
1. Acknowledge the Reality of Hurricanes knowledge of preferred alternatives based on
Accepting that hurricanes are natural existing relationships with stakeholders to
phenomena predicted to occur with higher determine the most optimal outcomes for
frequency and intensity is a necessary first all parties involved in the supply chain.
step in hurricane supply chain preparedness. 4. Effective Communication
With this understanding, it is optimal to form and Collaboration
a hurricane crisis committee that develops Each entity in the supply chain should
robust strategies to prevent breaks in the supply have dedicated hurricane crisis committees.
chain. These strategies should be tailored to Established lines of communication between
particular situations and involve regular risk the different entities’ committees are essential
assessments and updates to reflect the latest for mitigating the impact of hurricanes on
data and trends. supply chains.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 43

SUPPLY CHAIN RESILIENCE

In anticipation of any disruption, regular demands and allocate appropriate areas of

updates to all involved parties are essential to safety stock to address both customer needs
maintain order and clarity of plans. Members and the threat of storms.
of the hurricane crisis committee should 7. Ensure Insurance Coverage
receive information related to the event at Hurricane cargo insurance is a valuable
regular intervals and develop plans in response investment, a requirement for suppliers and
to events as they become predicted or known. an integral component of any hurricane
These regular updates to the committee should preparedness strategy. Regularly reviewing
be followed by equally regular and pertinent these insurance policies is paramount to
updates of the plan of action to involved maintaining adequate coverage against the
stakeholders, thus maintaining clear and myriad risks posed by hurricanes.
consistent communication of actions to meet 8. Dialogue and Capacity Discussions with
the needs of each involved entity. Often, there Receivers and Transportation Providers
may be communication breakdowns, so having With the disruption to supply chains
established messaging applications that can be during and in the aftermath of a hurricane,
accessed via cellular phone data or other forms it becomes necessary to coordinate with
of chat messaging allows for regular updates warehouses, receivers and transportation
and conversations with the different parties. providers to maintain the appropriate supply
With the These dedicated emergency communication to meet demand. These updates may require
plans streamline information flow to – and adjustments to the supplying warehouse(s),
disruption to facilitate responses from – the committee, modes of transportation or, in some cases,
supply chains and reassure stakeholders that plans are both. Regular communication with receivers
regularly updated and appropriately overseen. before these disruptions occur is important for
during and in 5. Prepare for Remote Operations understanding their operational capabilities
the aftermath Remote operations remain an important and constraints to accommodate alternate
aspect of risk mitigation strategies in hurricane modes of transportation, delivery schedules
of a hurricane, preparedness. Ensuring that key personnel and/or transit times. For example, in the
can operate remotely during hurricanes is aftermath of Hurricane Harvey, companies
it becomes crucial. Conducting phone tree exercises and with robust relationships in diversified
necessary to maintaining cloud-based data storage enhances transportation modes could effectively
operational continuity. This includes ensuring coordinate their inventory to their receivers
coordinate with that all critical systems and data are accessible via alternative and effective routes, even with
warehouses, remotely and that employees are trained to the disruption of multiple ports in Texas.
work effectively from remote locations. Determining existing relationships and
receivers and 6. Inventory Management capabilities for accessing different modal
transportation Safety stock is necessary for both supply deliveries, such as rail, intermodal and road,
chain disruptions and customer demand allows for preparation in deciding which
providers to changes. As hurricanes form and threaten alternatives to use in case of disruptions.
maintain the warehouses or distribution routes in one area,
it becomes essential to have readily available
Furthermore, knowledge of SKU-specific
safety stock locations and amounts in existing
appropriate warehousing sources and maintain safety stock warehouses unaffected by the hurricane
of appropriate SKU requirements in other disruption allows for meeting customer
supply to locations, especially during hurricane season. demand while preventing stockouts. With
meet demand. Companies that had similar strategic inventory such initial awareness, managing the logistics
management were best able to navigate of possible transport and existing warehouse
the U.S. Coast Guard’s Zulu declaration of relationships with appropriate SKUs can be
Port Tampa Bay and Port Canaveral during optimized for a rapid response in the case of an
Hurricane Helene, as illustrated later in this unpredictable disruption such as a hurricane.
article. Depending on various supply chain Generally speaking, before a hurricane, it
factors, including production, transportation is recommended to increase transportation
and warehousing capacity, it is reasonable to capacity within primary modes of transportation
temporarily increase safety inventories from to supply customers. Increased capacity
25% to 35% during hurricane season. The in primary transportation modes raises
safety stock inventory increase prior to the likelihood of receivers increasing
hurricane season and the subsequent their unloading capacity, thereby boosting
correction can be planned in advance, inventory levels. However, after a hurricane,
thus preventing a lack of warehousing and depending on various supply chain factors,
space or transportation capacity. it is generally observed that temporarily
Furthermore, having robust communication increasing modes of transportation with
with customers about their demand predictions faster transit time, such as trucks, may
allows for additional planning to meet these support the increase of inventory levels

44 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 following a disruption. These alternatives warehouses were best managed by companies
may incur different costs than initially with effective hurricane crisis committees and
anticipated. Logistics professionals must robust relationships with shipping companies,
be prepared to select the best alternative alternative ports, transportation companies and
options, monitor cost deviations and warehouses. For instance, vessels were redirected
provide timely visibility to stakeholders. to ports in Louisiana, such as the Port of New
9. Post-Hurricane Reflection and Orleans, and in Alabama, such as the Port of Mobile.
Continuous Improvement Response strategies included close coordination
With the conclusion of a hurricane event, it with receivers and transportation providers to
is crucial to identify areas for improvement restore inventory levels after hurricane-driven
in future risk mitigation strategies. After a closures using alternate modes of transportation
hurricane, the hurricane crisis committee while maintaining effective communication with
should reconvene to carefully review the receivers. Such careful consideration, combined
outcomes of the response strategies and with strong relationships, prevented stockouts
develop new plans based on the data. In and allowed for the maintenance of supply
addition to this response review, members chains in the face of such a wide-reaching
of the committee should be committed to hurricane impact.
exploring new technologies and strategies
that may assist with hurricane preparedness Conclusion
and resilience. Regular research into the latest Hurricane mitigation strategies in supply
logistics technologies, strategies and best chains require a robust combination of proactive
practices, as well as a review of prior efforts in planning and modeling by a dedicated committee,
hurricane management, will further enhance use of multiple technological tools, effective
the efforts of the hurricane crisis committee. communication lines and strategies, and a
commitment to regular improvement. By
Case Studies/Examples utilizing these operations, logistics professionals
The following case studies exemplify the practical can fortify the resilience of their supply chains
application of the risk-mitigation strategies in the face of the ever-increasing frequency
discussed earlier. These cases highlight the and intensity of hurricanes.
importance of proactive planning, inventory
DENNIS HERNANDEZ is a logistics manager at Suzano, managing
management, effective communication and robust logistics services procurement across North America. With more than a
coordination with the stakeholders involved. decade of experience, he has led initiatives to optimize operations and
enhance sustainability. Dennis is a Fulbright scholar and holds multiple
Hurricane Helene master’s degrees.
In September 2024, Hurricane Helene, a Category 4 GEMMA BERENGUER is an associate professor of management and a
storm, caused major disruptions in supply chains Ramón y Cajal Fellow at Universidad Carlos III de Madrid. She received
across the Southeastern United States [4]. The her Ph.D. from UC Berkeley’s IEOR Department in 2012. In 2023, she
U.S. Coast Guard’s response involved complete served as the president of the INFORMS Section on Public and Societal
Operations Research.
shutdowns of major ports under “Port Condition
Zulu.” Major ports such as Port Tampa Bay and REFERENCES
Port Canaveral experienced an effective halt on 1. National Oceanic and Atmospheric Administration (NOAA), 2023,
all inbound and outbound vessel traffic [5, 6]. “State of Science Fact Sheet: Atlantic Hurricanes and Climate
Companies with existing relationships with Change,” NOAA, May, https://sciencecouncil.noaa.gov/wp-content/
uploads/2023/05/1.1_SOS_Atlantic_Hurricanes_Climate.pdf.
alternative ports, such as those located in the West
2. Katharina Buchholz, 2024, “Tropical Cyclones Intensify Faster,”
Gulf (e.g., Port of Houston, Port of New Orleans) Statista, October 9, https://www.statista.com/chart/33216/
or mid- to North Atlantic (e.g., Port of New York/ number-of-atlantic-tropical-cyclones-undergoing--extreme--
New Jersey, Port of Baltimore), were best able rapid-intensification.
to optimally respond in the face of such a major 3. National Weather Service, 2024, “Hurricane Milton Impacts to
East Central Florida,” NOAA, https://www.weather.gov/mlb/
hurricane disruption, highlighting the importance
HurricaneMilton_Impacts.
of having a robust, proactive planning and inventory 4. National Weather Service, 2024, “Hurricane Helene: Record-
management strategy that allowed for alternative Breaking Rainfall and Historic Flooding,” NOAA, https://www.
sources of inventory to maintain supply. weather.gov/mrx/Hurricane_Helene.
5. Port of Tampa Bay, 2024, “Port Conditions Update - Hurricane
Helene,” September 28, https://www.porttb.com/2024/9/
Hurricane Harvey tropical-weather-update-tropical-cyclone-9.
In August 2017, Hurricane Harvey made landfall 6. Space Coast Daily, 2024, “U.S. Coast Guard Sets Port Condition ‘ZULU’
in Texas, creating disruptions at the ports Due to Hurricane Helene, Port Canaveral is Closed,” September 26,
along the coastline of the state [7]. Multiple https://spacecoastdaily.com/2024/09/u-s-coast-guard-sets-port-
vessels en route to these affected ports needed condition-zulu-due-to-hurricane-helene-port-canaveral-is-closed/.
7. Federal Emergency Management Agency (FEMA), 2017,
coordination with alternative ports to unload “Historic Disaster Response to Hurricane Harvey in Texas,”
their products. The efforts in rerouting vessels September 22, https://www.fema.gov/press-release/20230425/
and transporting goods from ports to appropriate historic-disaster-response-hurricane-harvey-texas.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 45

ROUNDTABLE PROFILE

QUANTUMBLACK’S HYBRID
APPROACH TO OPERATIONS
RESEARCH IN THE AGE OF AI
BY ALEX COSMAS, SOHRAB RAHIMI AND SARAH MULLIGAN

MODERN OPERATIONS RESEARCH (O.R.) solutions within defined constraints but have been
sits at a critical juncture. Traditional optimization limited by computational capacity and often struggle
methods that have defined the field for decades with highly dynamic, uncertain environments.
are now being transformed through integration AI is changing this paradigm in fundamental
with artificial intelligence (AI), creating new ways. Machine learning (ML) algorithms can now
possibilities – and new expectations – for solving detect patterns in vast datasets that would be
complex business problems. At QuantumBlack, invisible to traditional O.R. methods. They can
AI by McKinsey, we’re pioneering this integration adapt to changing conditions without explicit
with our “hybrid intelligence” approach, combining reprogramming and incorporate a broader range
the power of advanced AI with human expertise to of inputs, including unstructured data such as text,
revolutionize O.R. applications across industries. images and sensor readings.
However, AI alone isn’t the answer. Many systems
From Formula 1 to Enterprise Operations: operate as “black boxes,” making decisions that are
Our Data-Driven Heritage difficult to explain or validate. They may also fail to
QuantumBlack was born in the data-intensive world effectively incorporate domain-specific constraints
of Formula 1 racing, where milliseconds separate or business objectives. This is where our hybrid
victory from defeat and teams live and breathe data. approach shines, combining the interpretability
This heritage of extreme performance optimization and rigor of traditional O.R. with the adaptability
in complex, constrained environments has also and pattern recognition of AI.
played a significant role in shaping our approach to
operations research. As the AI arm of management OptimusAI: Reimagining
consultancy McKinsey & Company, we’ve built on Industrial Operations
this foundation by bringing together the precision of One of our flagship products, OptimusAI,
quantitative techniques with the strategic expertise exemplifies this hybrid approach. OptimusAI
from our consulting functions to help organizations is an industry-leading platform developed by
solve their most challenging operational problems. QuantumBlack Labs that transforms plant
At QuantumBlack, we bring together the foresight productivity and decision-making in materials,
and precision of data and technology with the energy and food production by integrating
creativity and understanding of people. The result traditional optimization techniques with
is hybrid intelligence – a source of competitive cutting-edge machine learning.
advantage that transforms how companies think, A recent deployment with a major copper
operate and disrupt. concentrate producer illustrates the power of this
approach. The company struggled with complex
The Evolution of O.R. in the AI Era operations involving thousands of sensors
Traditional operations research has long relied on feeding into control systems that had to solve
mathematical programming, simulation and other for nonlinear functions and multiple trade-offs.
analytical techniques to optimize complex systems. Traditional optimization approaches had reached
These methodologies excel at finding optimal their limits.

46 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 The QuantumBlack team deployed OptimusAI recommendations they receive. To address
to build an AI model that could use operating data this need, we’ve prioritized the development
to identify optimization opportunities for boosting of techniques for generating interpretable
copper production while maintaining quality. The insights from complex models, making the
system developed a proprietary control-room black box transparent.
advisor application that presented straightforward For example, in one manufacturing optimization
optimization recommendations to operators project, we created intuitive visualizations showing
despite working with thousands of variables how different production parameters contributed
behind the scenes. to quality outcomes. This helped operators
The results were remarkable: The plant improved understand not only what to adjust but why
throughput by 10%-15% and copper recovery by those adjustments mattered, which builds trust
2-4 percentage points. This represents millions in and facilitates adoption.
additional annual revenue with minimal capital
investment – a testament to the value of modernizing Applications in Industry
O.R. with AI. The true measure of this kind of hybrid approach
is the tangible impact it delivers across diverse
Merging O.R. and ML: Our industries. The following are some examples
Methodological Innovations of how we’re transforming O.R. in practice
There’s a tremendous amount of value and at QuantumBlack.
efficiency to be found in systematically bridging At QuantumBlack,
the methodological gap between traditional O.R. Process Manufacturing
and modern machine learning. Here are some key In industries such as mining, chemicals
we bring
aspects of the integrated approach we’ve taken. and energy, traditional approaches focus on together the
steady-state optimization, but real plants
Constraint-Aware Machine Learning operate in highly dynamic environments with foresight and
Traditional ML algorithms often struggle with frequent disruptions. precision
the hard constraints common in O.R. problems. By integrating reinforcement learning and
We’ve developed specialized architectures that digital twin technology with classical optimization, of data and
can incorporate operational constraints directly we’ve been able to create solutions that can respond technology with
into the learning process. For example, in supply to changing conditions and recover optimally
chain optimization, our models can learn efficient from disruptions. This approach helps optimize the creativity and
inventory policies while respecting warehouse performance in the context of thousands of
capacity constraints, service level agreements variables, allowing plants to improve throughput
understanding
and other business rules. while also becoming more energy efficient – of people. The
two objectives that are sometimes seen as
Reinforcement Learning for mutually exclusive. result is hybrid
Dynamic Optimization intelligence.
Many operational decisions involve sequential Healthcare
decision-making under uncertainty – a perfect The healthcare sector presents unique optimization
application for reinforcement learning (RL). challenges, with human lives at stake and highly
We’re using RL alongside traditional dynamic complex constraints. For one global pharmaceutical
programming to solve complex problems in company, we transformed their clinical trial
production scheduling and logistics routing. By operations using a combination of traditional
rewarding algorithms for decisions that optimize optimization techniques and machine learning.
long-term objectives rather than immediate We developed models to optimize country
gains, we’re achieving solutions that traditional footprint and site selection, incorporating
optimization methods would struggle to discover. predictive elements for enrollment periods
These models can adapt to changing conditions and risk factors. The result was a more than
in real time, continuously learning from new 10% increase in clinical trial productivity,
data to improve their recommendations. This is accelerating the time to market for critical
particularly valuable in volatile environments in medications while maintaining rigorous
which traditional static optimization models can quality standards.
quickly become outdated.
Retail
Explainable AI for Operations We can help companies move beyond siloed
One of the biggest challenges with AI-driven optimization to integrated supply chain
decision-making is transparency. Operations intelligence. For major retailers, we combine
managers need to understand and trust the traditional inventory optimization with demand

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 47

ROUNDTABLE PROFILE

forecasting using neural networks and The next frontier will involve even deeper
natural language processing to incorporate integration of these disciplines, with AI not only
unstructured data such as social media trends solving O.R. problems but also helping to frame them
and weather forecasts. more effectively. As more advanced techniques like
This approach allows for more accurate reinforcement learning and graph neural networks
predictions and optimal stocking decisions, mature, we expect to see O.R. applications that can
significantly reducing out-of-stock instances autonomously identify optimization opportunities
while simultaneously optimizing inventory and continuously adapt to changing conditions.
levels and reducing costs. For O.R. professionals, this represents both a
challenge and an opportunity. Those who embrace
Building Capabilities for the Future AI as a complement to their traditional toolset will
We recognize that implementing advanced AI/O.R. be positioned to deliver unprecedented value in the
hybrid solutions often requires organizations to years ahead. At QuantumBlack, we’re excited to be
develop new capabilities. This is why our approach at the forefront of this evolution, reimagining what’s
includes a specific focus on knowledge transfer possible when human expertise, data science and
As we look and capability building within client and partner artificial intelligence come together.
organizations. One advantage of being an AI
ahead, we see agency within a larger consultancy is that we can Editor’s note. The INFORMS Roundtable is the
recognized premier destination for top-level leaders
O.R. entering assemble multidisciplinary teams that combine
in OR/MS practice and delivers unique benefits to
various technical disciplines with industry-specific
a new age of expertise. These kinds of teams are well suited to member organizations, their representatives and
building capacity, designing training programs and the broader INFORMS community. To learn more,
accelerated supporting organizational shift with a wide range visit: https://tinyurl.com/mtchs4zf.
improvement, of clients.
ALEX COSMAS is a senior partner at QuantumBlack, AI by McKinsey,
We’ve also developed our own open-source
powered by the tools to support clients, partners and the wider
and leads AI for McKinsey in its travel and logistics sectors. He applies
the full data science toolkit to commercial and operational problems,
combination community in building AI-enhanced O.R. solutions. finding performance improvement opportunities through rapid model
Kedro, our open-source Python framework for development and deployment. A recognized expert in the use of predictive
of traditional building modular data science pipelines, has and probabilistic models to perform both deductive and inductive
reasoning from large datasets, Alex specializes in the areas of customer
optimization been donated to the Linux Foundation to help analytics, consumer choice, revenue management and pricing.
advance the field more broadly.
techniques As we look ahead, we see O.R. entering a new SOHRAB RAHIMI is a partner at QuantumBlack, AI by McKinsey, where he
helps organizations use AI to reinvent themselves and accelerate growth
with modern age of accelerated improvement, powered by the by creating new opportunities through reshaping their operations. Sohrab
combination of traditional optimization techniques works across industries, applying deep technical expertise in AI/GenAI
AI capabilities. with modern AI capabilities. This hybrid approach and optimization methods. His primary focus is on conversational and
will unlock previously impossible solutions to voice-enabled AI products, in which he has helped organizations enhance
complex operational challenges across industries, workforce management and operational excellence.
and QuantumBlack is committed to leading this SARAH MULLIGAN is director of marketing and communications at
transformation. By bringing together the rigor QuantumBlack, AI by McKinsey. She leads QuantumBlack’s global
and interpretability of traditional O.R. with the marketing and communications strategy that integrates storytelling with
adaptability and pattern recognition of AI, we’re data-driven insights to amplify McKinsey’s AI impact. Her role encompasses
corporate communications, brand strategy, demand generation and digital
helping organizations navigate complexity communications, focusing on translating complex technologies into
and make better decisions in an increasingly compelling narratives that drive business value. She has more than 15
dynamic world. years of experience spanning journalism and public relations.

48 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 “What makes data science a science is the fact that it interacts with the real
world. Find a customer. Universities teach you programming and math, but
it’s that interaction with someone who cares, that is very different and you
can’t teach it. Get a chance to work on a data problem with a real person –
that’s how you actually know if this is where you want to be in your career.”
Louis Luangkesorn, Lead Data Scientist, Highmark Health

“I love statistics. We’re living in an imperfect world and deal with a lot of
uncertainties and unknowns. Statistics is the perfect science to solve these
unknowns – I think it’s beautiful. I think it’s art.”
Xiaonan (Shannon) Shang, Manager of Data Analytics, Global Holdings and Adjunct Instructor,
University of Tulsa

“In the middle of COVID [pandemic] I said, ‘hey, what am I going to do

when I grow up?’ I really enjoy applying these types of methods to new
industries, so I thought about it and took the plunge in June 2021 with Data
Driven Supply Chain. It started out just me but now I’m very fortunate to
have a small team with me doing network design, inventory optimization,
forecasting, and a little bit of data visualization across sectors.”
Ralph Asher, Founder, Data Driven Supply Chain, and Adjunct Instructor at University of St. Thomas

For interviews and clips, scan QR code

or visit the INFORMS YouTube channel.
INFORMS INITIATIVES

FOSTERING ETHICAL BEHAVIOR IN

ACADEMIC PUBLISHING: A CALL
FOR SELF-REGULATION
BY CHRISTOPHER S. TANG AND MATTHEW WALLS

INFORMS HAS LONG BEEN A BEACON Understanding the Drivers of

for scholars and practitioners across various Unethical Behavior
disciplines, including management science, To promote self-regulation, it is essential to
operations research, marketing, strategy and understand why ethical breaches occur. Although
more. In 2024, the total submissions to all 17 most authors act in good faith, the academic
INFORMS journals exceeded 11,000, marking an environment is increasingly shaped by systemic
increase of more than 11% from 2023. This surge pressures that can distort incentives.
reflects robust engagement within the INFORMS “Publish or perish” dynamics tied to tenure,
community and reinforces the visibility and impact promotion or grant success can create an
of the organization’s scholarly platforms. overwhelming incentive to maximize publication
However, with this growth comes the challenge count. Institutional emphasis on metrics such as
of maintaining high ethical standards in publishing. journal impact factors and h-index scores further
Although most We propose that self-regulation within the INFORMS intensifies this drive. In global academic contexts
community can be a powerful and necessary tool to in which access to top-tier journals confers
authors act enforce ethical behavior and curb misconduct by a significant career advancement, the stakes
in good faith, small subset of authors. can be even higher.
These pressures, combined with gaps in training
the academic Upholding Ethical Standards or ambiguous standards across disciplines, can
environment As members of INFORMS, we each have a lead some authors – knowingly or otherwise –
responsibility to uphold high ethical standards on to engage in misconduct such as duplicate
is increasingly behalf of society, our institutions and the academic submissions, data manipulation or insufficient
profession. INFORMS’ ethics guidelines emphasize citation of related work.
shaped by honesty in reporting results, truthfulness in
systemic providing attribution and vigilance in identifying Measures to Promote Ethical Conduct
and discouraging actions that undermine the To reinforce ethical publishing, INFORMS
pressures that integrity of our field [1]. journals require submitting authors to
can distort • Honesty in Reporting Results: Authors acknowledge their understanding of the
must report research outcomes truthfully, organization’s ethical policies and guidelines.
incentives. regardless of whether the results support their Although slight variations exist across journals,
hypotheses. This principle safeguards the key principles are consistently emphasized:
credibility and cumulative value of research • Originality and Plagiarism: Authors must
in our field. clearly identify when portions of a manuscript
• Truthfulness in Attribution: Proper citation build upon previous work, including their own.
is essential when a manuscript draws on the Proper citation and an explanation of novel
ideas or language of prior work. This applies not contributions are expected.
only to work by others but also to authors’ own • Falsification and Fabrication: Altering or
previously published material. inventing data or results is unacceptable and
• Vigilance Against Unethical Actions: constitutes a serious ethical violation.
Community members are expected to remain • Redundant, Concurrent or Multiple
alert to behaviors that damage the profession Publications: Submitting similar or identical
and speak out when necessary. work to multiple journals – simultaneously or

50 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 sequentially – without disclosure is unethical, suspects that data has been fabricated or
as is fragmenting study results across numerous manipulated – or that similar papers have
publications without substantive differentiation. been submitted to multiple journals – they have a
responsibility to report these concerns. INFORMS
To support these policies, INFORMS journals editors treat such reports confidentially and
employ several safeguards. All co-authors are follow procedures designed to ensure
notified upon submission, helping to prevent fair resolution.
unauthorized or unilateral actions by any person. The mantra “See it, Say it, Sort it” captures
Additionally, a plagiarism detection system the spirit of self-regulation. When community
checks each manuscript against a wide corpus of members take an active role in maintaining
publicly available literature and other submissions high standards, they reinforce the expectation
within the INFORMS portfolio. of integrity and model ethical leadership We invite every
for others.
Ongoing Challenges
member of
Despite these measures, INFORMS continues to A Collective Responsibility the INFORMS
encounter ethical violations. Some submissions Ethical publishing is a shared enterprise.
closely mirror other uncited papers by overlapping Although formal policies and editorial oversight community to
author groups. In other cases, data has been are necessary, they are not sufficient on their embrace their
selectively manipulated at different stages of own. Ethical culture is built from the bottom
analysis to yield more favorable outcomes. Still, up – through daily choices, active mentorship role in fostering
others involve near-duplicate papers submitted to and a willingness to speak up when something an environment
multiple journals, often with different titles, within is amiss.
a short period. We invite every member of the INFORMS of trust,
Such actions not only violate INFORMS policy community to embrace their role in fostering an
but also erode trust in the academic publishing environment of trust, transparency and scholarly
transparency
system. Although journals have formal procedures rigor. By doing so, we not only protect the and scholarly
to investigate and respond to misconduct, editors reputation of our organization, but we also
cannot address these issues alone. ensure that the science we contribute to is rigor.
worthy of its purpose: to inform, to improve
The Crucial Role of Self-Regulation and to inspire.
This is where self-regulation becomes essential.
CHRISTOPHER S. TANG ([email protected]) is the Carter chaired
Self-regulation refers to the proactive role that professor at the UCLA Anderson School of Management. He is the
researchers take in upholding ethical norms – INFORMS VP of Publications.
monitoring their own practices, supporting ethical
behavior among colleagues and intervening when MATTHEW WALLS ([email protected]) is the director of publications
at INFORMS.
they observe misconduct.
If a co-author receives an email regarding a REFERENCE
submission they were not involved with, they should 1. https://www.informs.org/About-INFORMS/Governance/
promptly notify the editor-in-chief. If a researcher INFORMS-Ethics-Guidelines

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 51

INFORMS PRESIDENT-ELECT NOMINEES

RAJAN BATTA Management Science, Transportation Science and

Operations Research, and other premier journals,
e.g., European Journal of Operational Research.
Statement I am a Fellow of INFORMS (2022) and
AS PRESIDENT OF INFORMS, I WOULD have received the following awards from
focus on supporting the professional growth of INFORMS: Military and Security Society’s
both academic and industry members at various Koopman Prize (2024 and 2018) and Lifetime
stages of their development. For academic members, Achievement Award for Locational Analysis (2023).
this starts with supporting student chapters, I have also received many awards from other
students at the employment stage and training professional societies, most significantly, the Frank
opportunities for early-career faculty. For industry and Lillian Gilbreth Industrial Engineering Award
members, this is principally through creation of from IISE (2022).
networking opportunities. Our academic and
industry members can benefit from coordination Relevant Prior Experiences
I deeply value of industry/academic joint opportunities. I would • Leadership Positions: I have served in several
mentoring build on the excellent progress that INFORMS has leadership roles, including department chair,
made in the following areas: interim dean and associate dean in many
early-career • Networking: The greatest benefit of being different capacities. I have also served on the
faculty. I a member of INFORMS is expanding your board for IISE as Senior VP for Publications.
network. INFORMS should continue to develop • Experience in Faculty Mentoring and Peer
am further and fine-tune networking mechanisms Recognition: I currently serve as Associate
for members to connect with professional Dean for Faculty Affairs and Recognition. A
committed colleagues for mutual support and growth. significant part of my job is to provide support
to mentoring • INFORMS Student Chapters: INFORMS should for faculty mentoring and promote honors and
enhance its support by connecting chapters awards pathways.
doctoral with academics or industry professionals willing • Graduate Student Mentoring: I have supervised
students and to hold panels or talks. 58 doctoral dissertations and 57 master’s theses.
• Communities for New Students: INFORMS Currently, I have five doctoral and two master’s
supporting conferences should provide sessions for first- students in my lab.
them beyond year students to meet each other and get • Broad engagement with INFORMS: I have
assistance on where and how to get started served on the Chapters/Forums Committee,
their degree in the community. as president of ACORD, organizer of TutORials
• Career Resources: Expand and enhance in Operations Research at an INFORMS Annual
completion as the resume review/short interview slots Meeting, and an active member of several
they pursue at conferences and improve the quality of subdivisions and special interest groups.
INFORMS social media presence to attract • Industry Consulting Experiences: I have
diverse careers younger talent. extensive consulting experience in supply
in academia • Training Opportunities for Early-Career chain and manufacturing logistics. These
Faculty: INFORMS should invest and develop experiences have taught me to connect
and industry. programs that will specifically enhance training with analysts working in industry, and I can
in research, service and teaching. leverage this experience to help provide
• Joint Industry/Academic Opportunities: pathways for academics to connect with
INFORMS should provide a platform for industry practitioners.
speed-networking sessions to support the
synergy between industry and academia, using Bio
analytical tools that would benefit both sides. Rajan Batta, SUNY Distinguished Professor in the
Department of Industrial and Systems Engineering
Relevant Background at the University at Buffalo, focuses on logistical and
I am a SUNY Distinguished Professor in planning applications impacting services, such as
the Department of Industrial and Systems emergency response, disaster logistics and national
Engineering (ISE) at the University at Buffalo (UB). security. Rajan earned a B.Tech. from IIT, Delhi, and
I have a B.Tech. in mechanical engineering from Ph.D. in operations research from MIT. His research
the IIT, Delhi, and a Ph.D. in operations research spans the entire spectrum from highly applied to
from MIT. My work is in hazmat routing, location highly theoretical. He has supervised 58 doctoral
modeling, emergency response, and military dissertations; one-third of these graduates are
and security applications. I have published on a thriving in academia and two-thirds have excellent
regular basis in premier INFORMS journals, e.g., industry positions.

52 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 WEDAD J. brought faculty and students together with
leading practitioners. The goal was to enable

ELMAGHRABY
practitioners to inform research and teaching in
key areas such as Trustworthy AI, AI and Social
Inequity, Healthcare Analytics and Global Supply
Statement Chain Risk, while preparing students for analytics-
THANK YOU FOR THE OPPORTUNITY TO oriented jobs. Additionally, these initiatives helped
tell you about myself, my path, my actions and companies assess big data and AI opportunities
what I value. I am an engineer with a business and facilitate cutting-edge analytics and
mindset. I place tremendous value on industry technology dissemination.
connections, am committed to furthering Within INFORMS, I have served as president of
educational goals and appreciate INFORMS’ both the MSOM Society and Behavioral Operations
global stakeholder base. Management Society, VP of Conferences for WORMS,
If honored to serve as president, I will strengthen Secretary/Treasurer of the ENRE subdivision and a
I strive to foster INFORMS by increasing the visibility and impact member of the INFORMS Finance Committee, as
the society’s of operations research (O.R.) in solving real-world well as a subdivision liaison. I have also served as a
problems. Central to this is interdisciplinary and judge for prestigious award committees, including
growth by global collaboration: no single discipline, location or the Lanchester Prize, as well as chair of the George
uniting multiple professional track (academic, industry, government) B. Dantzig Dissertation Award. Through these
holds the silver bullet for today’s global challenges. engagements, I have gained a deep appreciation
stakeholders & We must foster strong connections to stay at the for our stakeholders’ priorities, diverse activities
forefront of technological advancements and their and the importance of evolving to meet members’
perspectives. O.R. applications. As we pursue this, we must ensure changing needs. As MSOM President, I initiated
My areas of continued education and workforce development the first Society Climate Survey to assess how we
opportunities to prepare future O.R. professionals could improve engagement and foster belonging
interest and and offer lifelong learning, connections and for all members.
growth for networking for members worldwide.
I hold the Dean’s Chair of Operations Bio
INFORMS: Management at the R. H. Smith School of Wedad J. Elmaghraby is the Dean’s Chair
visibility, Business, University of Maryland. I have been a of Operations Management at the Robert H.
faculty member in both engineering (Industrial Smith School of Business, University of Maryland,
connectivity, and Systems Engineering at Georgia Tech) and College Park. Her research interests are at the
business schools (Stern School of Business at NYU interface of operations management, economics
impact, and R. H. Smith School at UMD) for 25 years. My and behavioral decision-making. She currently
networking, collaborations with faculty across economics, serves as a department editor for Manufacturing &
marketing, information systems, statistics, public Service Operations Management, and Senior
upskilling, policy and engineering reflect a strong appreciation Associate Dean for Faculty at the Smith School
membership for industry applications and interdisciplinary work. of Business.
My work spans healthcare, energy, procurement, Wedad has been an INFORMS member for
growth and online market design, pricing and sustainability. 25+ years and has served as the president of both
support. My collaboration with companies and scholars the MSOM Society and Behavioral Operations
worldwide have proven the impact our community Management Society, VP of Conferences for
can have if we broaden our lens of value. Teaching WORMS, Secretary/Treasurer of the ENRE
engagements with universities in Southeast Asia, subdivision, member of the INFORMS Finance
Asia, the Middle East and Europe highlight the need Committee and subdivision liaison. She has served
to maintain strong global connections and meet as the director of the Deloitte Initiative on Artificial
each region’s entrepreneurial and diverse landscape. Intelligence and Learning and co-director of the
I served as co-director of the Smith Analytics Smith Analytics Consortium.
Consortium (SAC) and founder/director of the Wedad earned her B.S. and B.A. from Cornell
Deloitte Initiative in Artificial Intelligence and University in ORIE and economics, and M.S. and
Learning (DIAL) at Smith. Both initiatives Ph.D. from IEOR at UC Berkeley.

INFORMS 2025 ELECTION FOR THE 2026 BOARD OF DIRECTORS

To view the candidate profiles for the 2025 election, please visit https://www.informs.org/
About-INFORMS/Governance/INFORMS-Election/Slate-for-the-Board-of-Directors-Election

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 53

MEMBER INSIGHTS

THE META-PROBLEM: HOW TO CHOOSE

THE RIGHT PROBLEM TO SOLVE
BY ZOHAR STRINKA, CAP-X

WHEN IT COMES TO ANALYTICS, SOLVING Defining a Problem-Space

the right problem is great guidance, but it is also Comparing problems and recognizing how they
entirely unhelpful. Without an explanation of what may be similar or different is key to the notion of a
might make a problem right, there’s no way to know problem-space. It is also vital for making the concept
how best to act. of a problem a bit less slippery.
Borrowing from eye exams, I’ve found the best We can take an arbitrary problem that may be
way to recognize whether the problem is the wrong completely or incompletely defined, and craft a
problem to solve is through comparison. Just like problem-space by tinkering with as many aspects of
you can decide if lens A or B is a clearer picture, the problem we need. With math problems, the idea
you can decide if problem A or B is a better is straightforward: Change the numbers, operations
problem to solve. or steps. The smaller the changes, the closer your
Comparing problems is an abstract idea, but it is new problem is to the original.
We can take the first steppingstone to solving the meta-problem
an arbitrary of choosing which problem to solve. I like to use the Applying the Meta-Problem Approach
term problem-space to refer to the many different Choosing which problem to solve – i.e., solving a
problem problems from which we can choose. In this article, meta-problem – is a daily challenge. However, we
that may be I will define problems and the idea of a problem- don’t usually see it that way. Every time we pick
space, provide some examples, and illustrate how which task to complete on a to-do list or how to
completely or it can be beneficial to use this framework when treat a medical problem, we’re implicitly deciding
incompletely considering a practical dilemma. what we think is the best problem to solve.
Say you’ve gone to the doctor with a list of symptoms.
defined, Defining a Problem If the doctor concludes there’s no obvious cause for
The concept of a “problem” is at least as slippery as the issues, they may recommend waiting to see what
and craft a an eel. In school, we make it easier for students by will happen next. Effectively, they’ve decided that the
problem-space focusing on completely defined problems – ones problem to solve is that there are some unknowns,
that have a correct answer. One plus one equals and time is the best way to learn more.
by tinkering two. The derivative of x2 is 2x. An optimal solution But if the symptoms that brought you to the
with as many to a linear programming problem will always fall doctor are debilitating, “wait and see” may not be
on a corner point. a good choice because there are possibly more
aspects of the We usually aren’t so lucky in the real world. The urgent problems than the lack of information. The
problem right investment to make depends on unknowns. key is assessing the problem-space. Once you have
The right house to buy depends on preferences. reviewed the many versions of the problem in front
we need. The right problem to solve depends on trade-offs. of you, it’s easier to recognize which diagnostics or
Ultimately, the answer to problems we face as treatments may make the most sense.
adults will usually be “it depends.” Failing to take a broader view (by exploring the
Many problems are incomplete, so it makes problem-space) is where traditional problem-solving
sense to talk about choosing a problem in the techniques fall short. If the doctor says there isn’t
first place. Picking which house to buy solves the enough information to determine what’s wrong,
vague and open-ended problem of having a place it would be silly to argue that point if you are not
to live. Choosing a house that is a specific size, has a medical professional. But it would be entirely
good schools nearby or costs a certain amount is appropriate to share other problems that are in
somewhat more specific. Add enough criteria, and the problem-space that could be solved instead,
there is a right answer just like when asking what or suggest diagnostic tests to fill in the gaps in
one plus one equals. our knowledge.

54 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 There is another way in which traditional “Common sense” would say the problem she
approaches to problem-solving lead us astray. Once needed to solve was her homelessness, and buying
we have decided that “X” is the problem we need the candle was a waste of her scarce resources.
to solve, we tend to get locked in to “solutions But rebuilding her life was going to take effort and
mode.” We consider the question closed, and time, so she chose a better problem: how to find the
we stop learning. motivation and resilience to keep trying. The candle
In contrast, the meta-problem approach is provided that.
iterative. It asks us to choose a problem but keep How I frame and state the problem I want to
an open mind and be prepared to adapt. We solve has serious consequences in the real world.
don’t yet know what we don’t know, so as new Let’s take a business example. Suppose I’m the CEO
information comes to light, we continuously of a retail business, and I’m concerned that our
check to see if we need to pick a better problem. revenue is not growing. My marketing analytics
I will explore this idea further in the next article. show that our online store is not getting enough
traffic. I frame the problem as “we need more
Practical Examples visitors to our website,” and I invest time, talent and
I’ve found the best clue when deciding which millions of dollars in a set of marketing initiatives Failing to take
problem to solve is when something feels off. Maybe to attract visitors to our e-commerce website. a broader view
it’s just a gut feeling or a specific piece of evidence Let’s say it works. The next quarterly marketing
that the current problem is not ideal. In any of these report shows we almost doubled the number of (by exploring
cases, my goal is to lean into that discomfort to site visitors. However, very few of them bought the problem-
explore the problem-space further – but aimed at anything. We solved the wrong problem, and now
solving my concern. we still don’t have enough revenue and wasted a ton space) is where
Many years ago, I went to my doctor concerned of valuable resources.
about a foot injury. The doctor sent me to get
traditional
X-rays, which showed no issue, and called it a day. Conclusions problem-
But something didn’t feel right, so I eventually I believe the meta-problem approach offers a more
went back and asked to see a specialist. It took realistic way to pick the right problem to solve solving
a handful of doctor visits to get to the right one and helps avoid wasting limited resources solving techniques fall
(physical therapy), but at the end of the day, I got the wrong one. It prompts us to consider multiple
the treatment I needed. variations of the problem before we choose by short.
We already do this instinctively when something exploring the “problem-space” and asks us to
bothers us. But it’s a powerful tool to evaluate your keep revisiting our choice of problem as new data
alternative problems intentionally. Going back to emerges. Perhaps most importantly, the meta-
the eye exam scenario, trying on someone else’s problem method forces us to think very clearly
glasses rarely helps you see clearly. But if you test about what we are ultimately trying to achieve
several different lenses, you can quickly see which before we commit resources.
one fits you best. The next article in this series will focus on
A true story I read recently provides another a circular approach to problem-solving. The
good example: The writer had gone through a key takeaway is that when our problems have
harrowing period of homelessness, and at her unknowns, the only way to learn enough to find
lowest point, with little money and no shelter, she the best problem to solve is to iterate.
decided to purchase a luxury scented candle. She
kept the candle in her purse and used it to remind ZOHAR STRINKA, Ph.D., CAP-X, is a consultant focused on solving data
herself of the life she was trying to get back to. and process problems for her clients in a variety of industries.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 55

CONFERENCE PREVIEW

GET READY FOR THE 2025 INFORMS

INTERNATIONAL MEETING IN
SINGAPORE
BY JIE XU AND SZU HUI NG

WE ARE THRILLED TO WELCOME YOU • Networking receptions and informal

to the 2025 INFORMS International Meeting, special interest group meetups.
set to take place July 20-23 at the stunning • Special site tours in and around the
Raffles City Convention Centre in Singapore. Singapore area.
This international gathering brings together
thought leaders, researchers and practitioners Plenary and Keynote Speakers:
in operations research, analytics and the Voices Shaping the Future
management sciences from across the globe. We are proud to present an exceptional lineup
Whether you’re joining us to share your latest of plenary and keynote speakers whose insights
research, build new professional relationships span academia, industry and public policy. These
or explore innovative applications in decision distinguished leaders will frame the most pressing
science, this year’s conference promises to be challenges and opportunities facing the global
an unforgettable experience. analytics community:
• Yinyu Ye, visiting professor of Hong Kong
Scan below to Discover Singapore: Where University of Science and Technology, Shanghai
register today! Innovation Meets Culture Jiao Tong University and Chinese University
With its world-class infrastructure, commitment of Hong Kong at Shenzhen, will present several
to sustainability and fusion of global cultures, mathematical optimization problems/
Singapore is the ideal host city for this meeting. algorithms for AI such as large language
From the iconic Marina Bay skyline to hawker model training, tuning and inferencing.
centers teeming with culinary treasures, Singapore • Sheldon H. Jacobson, Founder Professor at the
offers attendees not only intellectual enrichment University of Illinois Urbana-Champaign, will
but also an immersive cultural experience. Be deliver a plenary talk on the future of aviation
sure to take some time to explore its botanical security and provide an overview of airport
gardens, waterfront promenades and security in the U.S. and how biometrics like
vibrant neighborhoods. facial recognition fit into the risk reduction
calculus for airport security.
What to Expect: Technical Sessions, • Teo Chung Piaw, Stephen Riady Professor at
Exhibits and More the NUS Business School and executive director
Over four days, attendees will have access to: of the Institute of Operations Research and
• Hundreds of technical presentations across Analytics (IORA) at the National University of
tracks including AI and data science, healthcare, Singapore (NUS), will present recent efforts
supply chain analytics, optimization, quantum to build theory-enabled, data-driven models
computing, semiconductor manufacturing for optimizing complex operational decisions
and more. across multiple sectors, with a focus on bridging
• Plenary and keynote talks featuring the gap between rigorous analytical frameworks
internationally renowned experts. and real-world implementation.

56 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

 • Jonathan Chang, vice president at VisionPower transit, shopping and cultural attractions.
Semiconductor Manufacturing Company Pte The venue offers modern amenities and
Ltd (VSMC), will offer an industry perspective spacious meeting rooms to support a full
on semiconductor manufacturing and explore program of technical sessions, exhibits
its revolution. and networking.
• Kelvin Bryan Tan, principal health economist INFORMS has secured special rates at the
and head of the Chief Health Economist headquarter hotel: Swissôtel The Stamford,
Office in the Ministry of Health, Singapore, which is attached to the convention center.
will give a keynote talk on operations Book your accommodations early to take
research for pandemic preparedness advantage of these rates and ensure proximity
in a post-COVID world. to the venue. There is a limited room block, and
• Stefan Woerner, principal research scientist discounted rates will expire June 27 – and rooms
and manager of the Quantum Computational are filling fast! Visit: https://meetings.informs.org/
Science group of IBM Quantum at IBM Research wordpress/2025international/venue.
Europe – Zurich, will discuss quantum computing
as the new computational paradigm with the Travel and Visa Information Book your
potential to transform many disciplines. Singapore’s Changi Airport (SIN) is one of
• Ming Hu, Distinguished Professor of Business the world’s most connected hubs, with easy access accommodations
Operations and Analytics at the University of to the city center by mass rapid transit (MRT) early to take
Toronto, will share recent work in building or taxi.
models and theories for supply and demand Attendees traveling from abroad should ensure: advantage of
matching over a space. • Passports are valid for at least six months from hotel rates and
• Sharon Ng, professor and deputy dean at your date of entry.
Nanyang Business School, will advocate for • You review visa requirements for your ensure proximity
bridging the divide to advance knowledge country via the Singapore Immigration &
through industry-academia collaboration. Checkpoints Authority.
to the venue.
• You complete the mandatory SG Arrival There is a limited
This year’s plenary and keynote sessions are Card (an electronic document for all visitors)
designed to spark dialogue, expand perspectives before arrival. room block –
and ignite fresh ideas. We encourage all attendees discounted rates
to attend these featured talks to gain strategic Visit the Travel page for more: https://meetings.
insight into where our field is headed. informs.org/wordpress/2025international/venue. will expire
June 27!
Registration Details Be Part of Something Global
Early registration is open now through June 16. From fostering new collaborations to engaging with
Secure your place at discounted rates: cutting-edge research and policy developments,
• INFORMS members: $730 the 2025 INFORMS International Meeting is
• Nonmembers: $1,010 your opportunity to be at the forefront of
• Student/retired members: $300 analytics innovation.
• Student/retired nonmembers: $420 On behalf of the organizing committee, we invite
you to join us in Singapore this July. Let’s exchange
After June 16, standard rates will apply. Please ideas, spark new partnerships and shape the future
note that all presenters, session chairs and attendees of our discipline – together.
must register and pay the appropriate fee. One-day We can’t wait to welcome you.
passes and discounted options are not available.
To register, visit: https://meetings.informs.org/ JIE XU is a professor of systems engineering and operations research at
wordpress/2025international/register. George Mason University in Fairfax, Virginia. He is general co-chair of the
2025 INFORMS International Meeting.

Venue and Accommodations

SZU HUI NG is associate professor and head of the Department of
The 2025 INFORMS International Meeting will Industrial Systems Engineering and Management at the National
be held at the Raffles City Convention Centre, University of Singapore. She is general co-chair of the 2025 INFORMS
centrally located with convenient access to International Meeting.
INFORMS AWARDS & PRIZES

2025 EDELMAN GALA

CELEBRATES EXCELLENCE
The 2025 INFORMS Analytics+ Conference was held on April 6–8 in Indianapolis, with the annual
Edelman Gala showcasing the best of the best in analytics. The full video of the 2025 Edelman Gala
and award ceremony can be viewed on INFORMS YouTube channel. The following awards and winners
were announced during the gala.

2025 FRANZ EDELMAN AWARD

The prestigious 2025 Edelman Award was
bestowed to USA Cycling for its use of
operations research (O.R.) and analytics in
elite sports performance. With Project 4:05,
USA Cycling transformed race strategy and
athlete preparation, enabling the Women’s
Team Pursuit to win gold at the 2024
Paris Olympics.

2025 INFORMS PRIZE

Micron Technology, Inc., a global leader in
memory and storage solutions, was awarded
the 2025 INFORMS Prize. This prestigious
award recognizes Micron’s use of advanced
analytics and O.R. to optimize its end-to-end
supply chain to deliver best-in-class reliability,
responsiveness and cost.

2025 UPS GEORGE D. SMITH PRIZE

The UPS George D. Smith Prize recognizes
excellence in preparing students to become
practitioners of operations research (O.R.)
and analytics. The 2025 winner is the
National University of Singapore (NUS),
Master of Science in Business Analytics.
The award recognizes excellence in
preparing students to become
practitioners of O.R. and analytics.

58 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

2024 WAGNER PRIZE
The 2024 Daniel H. Wagner Prize for
Excellence in the Practice of Advanced
Analytics and Operations Research was
awarded to a team from JD.com for
improving fulfillment efficiency with data-
driven integrated assortment planning
and inventory allocation. The award was
presented at the 2024 INFORMS Annual
Meeting in Seattle.

2025 EARLY CAREER

PRACTITIONER AWARD
The 2025 INFORMS Early Career Practitioner
Award was presented to Lili Zhang, research
engineer at Hewlett Packard Enterprise (HPE)
for revolutionizing AI-driven decision-making
at HPE by leveraging machine learning,
knowledge graphs and advanced analytics
to enhance digital marketing, customer
engagement and strategic business growth.

2025 INNOVATIVE APPLICATIONS

IN ANALYTICS AWARD (IAAA)
The 2025 Innovative Applications in Analytics
Award (IAAA) was awarded to a team from
the Ford Motor Company, for their project
“Optimizing Product Feature Offering
through Connected Vehicles Analytics,
Intelligent Sampling, and Generative AI,”
which was announced during a special
luncheon during the 2025 INFORMS
Analytics+ Conference.

For more information on these prizes, please visit https://www.informs.org/Recognizing-Excellence/INFORMS-Prizes.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 59

STUDENT PERSPECTIVES

DRIVING AUTONOMY: THE CRITICAL

ROLE OF OPERATIONS RESEARCH
AND ARTIFICIAL INTELLIGENCE
BY PAVITHRA SRIPATHANALLUR MURALI

AUTONOMOUS SYSTEMS ARE RESHAPING support industries and economies in addressing

the world as we know it in terms of climate various issues worldwide. Globally, it is estimated
management, warehousing, medicine and that around 60% of companies will adopt these
logistics. The course of history has taught us autonomous systems by 2025 [2]. This deployment
how technological advancements alter the way is expected to improve decision-making processes,
economies operate and how people or society safety for human activities and reliability of
coexist. The Industrial Revolution transformed production. The manufacturing industry is
manufacturing and labor. The digital age redefined already highly automated and continues to rise.
access to information and communication. Today, Beyond manufacturing, fleets of driverless
we stand at the frontier of another shift led by autonomous vehicles are an apt example of how
autonomous systems that are reshaping how we these modern technologies affect our everyday
think about logistics, mobility, medicine, climate, lives. Self-driving cars in the streets of Phoenix, San
infrastructure management and beyond. Francisco and Tokyo are operating as robotaxis
providing ride-share services to local citizens [3].
From Automation to Autonomy, Why Now? Autonomous vehicles bring benefits in terms
Automation can be described as an enabler of a of reducing congestion and fuel emissions and
system governed by a set of rules that essentially improving safety and travel times. However,
control that system. For example, a thermostat is integrating them in existing public transportation
a good example of automation in which a specific infrastructure requires crucial study and assessment.
state of home climate is controlled by a specific set Similarly, fleets of robots serving a geographical
of rules. Autonomy, on the other hand, effectively area can be tasked to address various needs such
introduces uncertainty into this process, in which as delivering medicine and food in remote areas,
the system doesn’t have a clear set of rules but rather tackling wildfires and monitoring vegetation. Their
has guidelines that help navigate uncertainty while applications have no limit and are customizable for
making decisions to overcome new uncertain many tasks. Fleets of specialized robots acting in
scenarios. Provided with a map of information, coordination within enclosed areas like warehouses
autonomous systems use data and algorithms to or other logistical locations carry out tedious tasks
make decisions following intelligent reasoning. In of moving packages and organizing inventory in
other words, autonomous systems operate with a highly efficient manner. Similarly, robotic arm
a knowledge-based decision-making approach systems aid in packing packages in an optimized
that is more complex than the if-then use cases manner to efficiently utilize space in pallets.
provided via training. Enabling autonomy involves
integrating existing systems with sensors and Tackling Uncertainty:
algorithms. As such, physical autonomous systems The Role of O.R. and AI
are a conglomeration of mechanical systems, The question now arises: How do autonomous
sensor technology, connectivity and artificial systems manage the uncertainty to which they
intelligence (AI). Recently, these systems have are exposed in the real world? Unlike traditional
been at the forefront of development owing to automated systems – working in controlled and
their promise and value in meeting the sustainable predictable worlds – autonomous systems have
development goals stipulated by the United to decide on the basis of dynamic worlds full of
Nations, which address equitable access to new scenarios, interruptions and incomplete
all necessities to make livelihoods safer, more information. Whether it’s an autonomous car
resilient and sustainable for humans [1]. changing course in response to an unforeseen
roadblock or a delivery drone rerouting because of
Autonomous Systems at Scale: Enabling inclement weather, these systems need more than
Intelligence Across Industries preprogrammed rules – they need smart structures
Autonomous systems operating in fleets, such that enable them to reason, adjust and respond in
as connected autonomous vehicles, drones ambiguity. This is where the fields of operations
and robots, provide strategic intelligence and research (O.R.) and AI play a critical role. They

60 ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2

complement each other to offer the mathematical environments [9]. Finally, multirobot coordination
modeling, optimization and learning that allow within logistics networks is facilitated through
autonomous systems to make real-time and combined modeling techniques. Storage location
at-scale decisions, as well as systems to reason assignment and path planning are integrated
in uncertainty, optimize performance and learn into one optimization problem, resulting in
from the environment. more throughput and better energy efficiency
Recent advancements in agentic AI reflect by coordinating physical layout decisions and
the rapidly increasing pace toward autonomous routing logic [10].
intelligence, which essentially propels the ability Together, these studies illustrate the ways
to achieve complex, long-horizon objectives in which AI and O.R. enable decision-making
independently. These systems are designed to in autonomous systems and provide the tools
operate with minimal guidance and navigate a necessary to deal with uncertainty. From
vast range of situations without being explicitly transportation and warehousing to planning
programmed [4]. The development of such intelligent and system robustness, these techniques are
agents is the path toward autonomy that cannot constructing the future of autonomy for the real
merely respond but also plan, learn and improve world. With autonomous systems transitioning Recent
behavior over time. across domains, the synergy of O.R. and AI will
Simulation-based approaches are now becoming increasingly determine their effectiveness. From advancements
vital in characterizing emergent behavior in complex managing real-time fleet management to learning in agentic
multiagent systems. For instance, studies on viable policies in random settings, these systems
the resilience of connected and autonomous are based on solving core problems related to AI reflect
vehicles under cyber-physical disruptions have adaptability, scalability and ethical alignment.
demonstrated how agent-based simulation assists Decentralized decision-making increases the
the rapidly
in understanding the impact of disruptions on the flexibility and robustness of autonomous fleets, increasing
travel time under vehicle behavior dynamics and especially where communication latency, localized
redundancy mechanisms [5]. The simulation models disruption or scale renders central control infeasible. pace toward
facilitate policy analysis, safety evaluation and more Further, it helps distributed agents to make autonomous
resilient strategy design. autonomous, context-dependent decisions in
In urban transport, O.R. techniques have been pursuit of common or different objectives. intelligence,
applied to deal with uncertainty of demand. For Future work must focus on ensuring the which
example, researchers consider optimizing fleet size development of algorithms that prioritize safety
and operation of autonomous and conventional and transparency, which run efficiently in low- essentially
buses. Using a stochastic optimization, the authors resource settings and are highly resilient to
demonstrate that adopting autonomous buses dynamic environments. As autonomy expands
propels
enables more flexible routes and lower total into socially important areas – transportation, the ability
cost with service quality preserved in uncertain healthcare, logistics and climate resilience –
demand [6]. Similarly, the optimization of chargers building transparency, fairness and governance to achieve
for fleets of battery electric buses remains a into their design is crucial. In this new world, complex,
challenging task. One study suggests an integrated O.R. and AI are not optimization tools in isolation;
optimization model accounting for charger they are the drivers of smart, adaptive and long-horizon
placement and scheduling under opportunity decentralized decision-making that can enable objectives
charging. The outcome is a more efficient system reliable, resilient and sustainable systems for
with ensured operational feasibility coupled with the future. independently.
reduced total energy and infrastructure costs [7].
In industrial environments such as warehouses, Note. The reference list can be found online:
deep reinforcement learning (DRL) has been used https://doi.org/10.1287/orms.2025.02.04.
to effectively solve high-dimensional decision
PAVITHRA SRIPATHANALLUR MURALI is a fourth-year Ph.D. candidate
space problems. A study demonstrates how DRL
in systems engineering and operations research at George Mason
can be used to create a robust robotic packing University. Her research is focused on simulation and reinforcement
system that learns optimal solutions for 3D learning, with particular emphasis on networked systems such as
bin-packing problems – adapting to different infrastructure, supply chain and logistics. Her work explores innovative
object sizes and constraints without hardcoding approaches to enhance the safety, efficiency and resilience of these
critical systems. She is an editorial staff writer for OR/MS Tomorrow,
each possible situation [8]. Navigation under
the student magazine of INFORMS.
uncertainty is another traditional problem area
in which metaheuristic optimization techniques
have proven fruitful. The grey wolf optimization
algorithm was customized and used for the
autonomous robot path-planning problem, in
which it assists in finding cost-effective and
obstacle-free paths, even in unknown or dynamic
 THANK YOU TO OUR
#2025ANALYTICS SPONSORS!

PLATINUM LEVEL

SILVER LEVEL BRONZE LEVEL

COLLEGE OF COMPUTING AND

SOFTWARE ENGINEERING
School of Data Science and Analytics

DAY TWO KEYNOTE

ANALYTICS SOCIETY

$10,000 ANALYTICS+ POSTER COMPETITION

UPS GEORGE D. SMITH PRIZE

EDELMAN PREMIER

EDELMAN EXECUTIVE

EDELMAN CHAMPION
LAST WORD

TECHNOLOGY MODERNIZATION
GONE WRONG: THE COST OF IGNORING
CUSTOMER EXPERIENCE
BY POOJA DEWAN

MODERNIZATION SHOULD ENHANCE THE I’ve seen this happen in the corporate
customer experience, not erase what makes it world too often. A company embarks on a bold
special. Yet, a recent change at my long-time yoga modernization effort, pouring millions into new
studio has turned an exciting transformation into technology, but in the process, they unintentionally
a cautionary tale. disrupt what customers value most.

A Renovation That Broke More Than It Fixed What the Studio Could Have
This studio wasn’t just a place to exercise; it was Done Differently
my sanctuary. It saw me through the early stages Modernization doesn’t mean abandoning what
of motherhood, the pressures of an executive works. Instead of replacing instructors, the studio
career, the challenges of COVID-19 and even could have:
the devastation of losing my home to a fire. The • Offered hybrid classes: Combined in-person
instructors were skilled, and the mix of people – and virtual options, cutting labor costs and
from beginners to experts – created a community potentially adding another value to attract a
that kept me coming back. different customer base.
It wasn’t perfect. The carpets were worn, and the • Enhanced, not replaced, human connection:
HVAC system broke down at times. But none of that Used technology to complement live instruction,
mattered because the experience was rich. Even not eliminate it. Having deployed so many
when membership fees increased after the COVID-19 Internet of Things (IoT) solutions, I could
pandemic, we remained loyal. The studio wasn’t just think of so many ways we could have used
a gym – it was home. artificial intelligence (AI) and sensors to
So, when they announced a renovation, we create cost-effective solutions to cutting
were thrilled. New floors, mirrors and equipment labor costs.
felt like a well-earned upgrade. But what came • Listened to customers first: Understood what
next was perplexing. matters before making sweeping changes.
Instead of improving the experience, the studio
replaced its instructors with prerecorded videos. These same principles apply to enterprise
The energy, connection and sense of belonging – technology transformations.
gone. A space that once buzzed with life was now
eerily empty. A Business Parallel: When AI and
Data Modernization Go Wrong
When Modernization Misses the Mark This experience reminded me of a project in
In an effort to cut labor costs and modernize which I helped a digital leader develop data
infrastructure, the studio forgot what made modernization and AI strategy. The company
it successful. The result? Over 95% of loyal faced massive infrastructure costs, but as we
customers left. dug deeper, we realized something bigger: Their
Businesses often fall into this trap. They focus outdated tech wasn’t just expensive – it was
on cost-cutting and technology upgrades, actively blocking growth. Their fragmented
forgetting that customers stay for the experience – data systems slowed decision-making, created
not the infrastructure. poor customer experiences and hurt revenue.

ORMSTODAY.INFORMS.ORG I 2025, ISSUE 2 63

LAST WORD

If we had focused purely on infrastructure In the race for AI and digital transformation,
cost reduction, we would have missed the real success doesn’t come from flashy pilots or cutting-
opportunity of aligning technology, processes edge tools alone. It comes from creating real value
and people to fuel business growth. Instead, for customers.
we streamlined their systems, enabling faster,
more personalized customer interactions. The Urgency to Act Now
The result? Cost savings, efficiency gains and As you embark on your next technology
increased sales. modernization project, ask yourself:
• Are we upgrading to solve real business
The Lesson for Modernization challenges, or just for the sake of it?
Modernization should be a growth enabler – • How will this impact the customer
not just a cost-cutting exercise. experience, both short- and long-term?
The yoga studio’s failure serves as a • Are we balancing cost savings with investments
powerful reminder: If you lose sight of the that create value for customers?
customer experience, you lose the customer.
The same is true for any transformation Because in the end, your customers – not your
of technology. costs – will determine your success.
Too many companies focus on technology first, The time to rethink modernization is now. Let’s
business second and customers last. But success connect and discuss how to align technology,
comes from flipping that equation: processes and talent to drive real, sustainable value –
1. Customers first – How does this impact for both your business and your customers.
their experience?
2. Business value second – Will this drive POOJA DEWAN is a technology and data executive with deep experience
driving innovation and digital transformation in Fortune 200 companies.
growth, efficiency or both? She has held executive roles including Chief Data & Analytics Officer
3. Technology last – What’s the right tech to at Otis and Chief Data Scientist at BNSF Railway. Pooja currently advises
support this strategy? organizations on data strategy, AI adoption and scaling innovation responsibly.

SCAN QR CODE TO VISIT RESOUNDINGLYHUMAN.COM

Subscribe today at Amazon Music, Apple Podcasts, Google Podcasts & Spotify!
 REGISTER TODAY!
#INFORMS2025
Don’t miss the 2025 INFORMS Annual Meeting in Atlanta, where more than 6,000
INFORMS members, students, prospective employers and employees, and academic
and industry experts will share the ways operations research (O.R.) and analytics are
fueling smarter decisions for a better world.

From dynamic keynote speakers to interactive workshops and numerous career and
networking opportunities, the 2025 Annual Meeting promises to ignite inspiration
and foster collaboration, allowing the brightest minds in O.R. and analytics to shine.

Join INFORMS on October 26-29, in Atlanta, where we will unite in our quest for
smarter decisions, forging a path toward a better world for generations to come.

Speaker Registration Deadline: August 1

informsannual.com

INFORMS is more than a meeting, it is your forever professional home.