Skip to main content

Alma Rahat

Swansea University, Computer Science, Faculty Member

Followers

45

Following

44

Co-author

1

Public Views

I am a Senior Lecturer in Big Data/Data Science at Swansea University, UK. I have a BEng (Hons) in Electronic Engineering from the University of Southampton, UK, and a PhD in Computer Science from the University of Exeter, UK. I worked as a product development engineer after my bachelor's degree, and held post-doctoral research positions at the University of Exeter. Before moving to Swansea, I was a Lecturer in Computer Science at the University of Plymouth. My current research focus is in the broad areas of fast hybrid optimisation methods, real-world problems and machine learning. In particular, I am developing efficient methods inspired from surrogate-assisted (Bayesian) optimisation for optimising computationally or financially expensive problems (for example, computational fluid dynamics design problems).

less

Richard Everson

Gilberto Rivera

Universidad Autonoma de Cd. Juarez

IIT Roorkee

Nenad Mladenovic

Shengxiang Yang

De Montfort University, Leicester, UK)

Centro de Investigacion y Estudios Avanzados del IPN

Interests

Uploads

Papers by Alma Rahat

Proceedings of the Genetic and Evolutionary Computation Conference

Optimisation problems often have multiple conflicting objectives that can be computationally and/... more Optimisation problems often have multiple conflicting objectives that can be computationally and/or financially expensive. Monosurrogate Bayesian optimisation (BO) is a popular model-based approach for optimising such black-box functions. It combines objective values via scalarisation and builds a Gaussian process (GP) surrogate of the scalarised values. The location which maximises a cheap-to-query acquisition function is chosen as the next location to expensively evaluate. While BO is an effective strategy, the use of GPs is limiting. Their performance decreases as the problem input dimensionality increases, and their computational complexity scales cubically with the amount of data. To address these limitations, we extend previous work on BO by density-ratio estimation (BORE) to the multi-objective setting. BORE links the computation of the probability of improvement acquisition function to that of probabilistic classification. This enables the use of state-of-the-art classifiers in a BO-like framework. In this work we present MBORE: multi-objective Bayesian optimisation by densityratio estimation, and compare it to BO across a range of synthetic and real-world benchmarks. We find that MBORE performs as well as or better than BO on a wide variety of problems, and that it outperforms BO on high-dimensional and real-world problems. CCS CONCEPTS • Computing methodologies → Modeling methodologies; • Theory of computation → Gaussian processes; • Applied computing → Multi-criterion optimization and decision-making.

On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints

Lecture Notes in Computer Science, Jul 19, 2020

We are often interested in identifying the feasible subset of a decision space under multiple con... more We are often interested in identifying the feasible subset of a decision space under multiple constraints to permit effective design exploration. If determining feasibility required computationally expensive simulations, the cost of exploration would be prohibitive. Bayesian search is data-efficient for such problems: starting from a small dataset, the central concept is to use Bayesian models of constraints with an acquisition function to locate promising solutions that may improve predictions of feasibility when the dataset is augmented. At the end of this sequential active learning approach with a limited number of expensive evaluations, the models can accurately predict the feasibility of any solution obviating the need for full simulations. In this paper, we propose a novel acquisition function that combines the probability that a solution lies at the boundary between feasible and infeasible spaces (representing exploitation) and the entropy in predictions (representing exploration). Experiments confirmed the efficacy of the proposed function.

$Research paper thumbnail of $\epsilon$-shotgun: $\epsilon$-greedy Batch Bayesian Optimisation$

$\epsilon$-shotgun: $\epsilon$-greedy Batch Bayesian Optimisation

arXiv: Learning, Feb 5, 2020

Bayesian Search for Robust Optima

Many expensive black-box optimisation problems are sensitive to their inputs. In these problems i... more Many expensive black-box optimisation problems are sensitive to their inputs. In these problems it makes more sense to locate a region of good designs, than a single—possibly fragile—optimal design. Expensive black-box functions can be optimised effectively with Bayesian optimisation, where a Gaussian process is a popular choice as a prior over the expensive function. We propose a method for robust optimisation using Bayesian optimisation to find a region of design space in which the expensive function’s performance is relatively insensitive to the inputs whilst retaining a good quality. This is achieved by sampling realisations from a Gaussian process that is modelling the expensive function, and evaluating the improvement for each realisation. The expectation of these improvements can be optimised cheaply with an evolutionary algorithm to determine the next location at which to evaluate the expensive function. We describe an efficient process to locate the optimum expected improve...

seanwalton/mixed-initiative-procedural-dungeon-designer: First Release

A unity3d tool which designs dungeon maps based on a level designer's preferences

Shape optimisation of the sharp-heeled Kaplan draft tube: Performance evaluation using Computational Fluid Dynamics

Renewable Energy, 2020

In the first of this two-part contribution, a methodology to assess the performance of an elbow-t... more In the first of this two-part contribution, a methodology to assess the performance of an elbow-type draft tube is outlined. using Computational Fluid Dynamics (CFD) to evaluate the pressure recovery and mechanical energy losses along a draft tube design, while using open-source and commercial software to parameterise and regenerate the geometry and CFD grid. An initial validation study of the elbow-type draft tube is carried out, focusing on the grid-regeneration methodology, and the use of a steady-state assumption for evaluating the design's efficiency. The Grid Convergence Index (GCI) technique was used to assess the uncertainty of the pressure recovery to the grid resolution. It was found that estimating the pressure recovery through area-weighted averaging significantly reduced the uncertainty due to the grid. Simultaneously, it was found that this uncertainty fluctuated with the local cross-sectional area along the geometry. Subsequently, a study of the inflow cone and outer-heel designs on the flowfield and pressure recovery was carried out. Catmull-Rom splines were used to parameterise these components, so as to recreate a number of proposed designs from the literature. GCI analysis is also applied to these designs, demonstrating the robustness of the grid-regeneration methodology.

An interactive tool for enhancing hospital capacity predictions using an epidemiological model

Proceedings of the Genetic and Evolutionary Computation Conference Companion, 2021

Hospital managers have limited resources, and they have to plan on how best to allocate these res... more Hospital managers have limited resources, and they have to plan on how best to allocate these resources based on predicted demand, which is often based on linear or exponential models. In this paper, we propose an interactive tool that produces a forecast of bed occupancy based on an epidemiological model. We optimise this model to fit recently observed data, and interactivity is conferred through a controllable parameter of the model such that users can readily investigate hypothetical scenarios for planning purposes. This study was designed with the Welsh National Health Service, and was born out of their practical need of accurately modelling hospital occupancy during the ongoing Covid-19 pandemic.

Surrogate-Assisted Evolutionary Optimisation (SAEOpt'16) Chairs' Welcome & Organization

Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion - GECCO '16 Companion, 2016

It is our great pleasure to welcome you to the Workshop on Surrogate-Assisted Evolutionary Optimi... more It is our great pleasure to welcome you to the Workshop on Surrogate-Assisted Evolutionary Optimisation (SAEOpt). In many real world optimisation problems evaluating the objective function(s) is computationally expensive. Surrogate-assisted optimisation attempts to alleviate this problem by employing computationally cheap 'surrogate' models to estimate the objective function(s) or the ranking relationships between candidate solutions. Surrogate-assisted approaches have been widely used across the field of evolutionary optimisation, and successful applications include aerodynamic design optimisation, structural design optimisation, data-driven optimisation, chip design, drug design, robotics and many more. Despite recent successes in using surrogate-assisted evolutionary optimisation, there remain many challenges. In this workshop we aim to promote research on surrogate-assisted evolutionary optimisation, particularly the synergies between evolutionary optimisation and machine learning. We hope that the workshop will be of interest to researchers working on the applied and theoretical aspects of surrogate-assisted evolutionary optimisation.

Hybrid Evolutionary Approaches to Maximum Lifetime Routing and Energy Efficiency in Sensor Mesh Networks

Evolutionary Computation, 2015

Mesh network topologies are becoming increasingly popular in battery-powered wireless sensor netw... more Mesh network topologies are becoming increasingly popular in battery-powered wireless sensor networks, primarily because of the extension of network range. However, multihop mesh networks suffer from higher energy costs, and the routing strategy employed directly affects the lifetime of nodes with limited energy resources. Hence when planning routes there are trade-offs to be considered between individual and system-wide battery lifetimes. We present a multiobjective routing optimisation approach using hybrid evolutionary algorithms to approximate the optimal trade-off between the minimum lifetime and the average lifetime of nodes in the network. In order to accomplish this combinatorial optimisation rapidly, our approach prunes the search space using k-shortest path pruning and a graph reduction method that finds candidate routes promoting long minimum lifetimes. When arbitrarily many routes from a node to the base station are permitted, optimal routes may be found as the solution ...

Multi-objective routing optimisation for battery-powered wireless sensor mesh networks

Mesh network topologies are becoming increasingly popular in battery powered wireless sensor netw... more Mesh network topologies are becoming increasingly popular in battery powered wireless sensor networks, primarily due to the extension of network range and resilience against routing failures. However, multi-hop mesh networks suffer from higher energy costs, and the routing strategy directly affects the lifetime of nodes with limited energy sources. Hence while planning routes there are trade-offs to be considered between individual and system-wide battery lifetimes. We present a novel multi-objective routing optimisation approach using evolutionary algorithms to approximate the optimal trade-off between minimum lifetime and the average lifetime of nodes in the network. In order to accomplish this combinatorial optimisation rapidly and thus permit dynamic optimisation for self-healing networks, our approach uses novel k-shortest paths based search space pruning in conjunction with a new edge metric, which associates the energy cost at a pair of nodes with the link between them. We demonstrate our solution on a real network, deployed in the Victoria & Albert Museum, London. We show that this approach provides better trade-off solutions in comparison to the minimum energy option, and how a combination of solutions over the lifetime of the network can enhance the overall minimum lifetime.

On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints

We are often interested in identifying the feasible subset of a decision space under multiple con... more We are often interested in identifying the feasible subset of a decision space under multiple constraints to permit effective design exploration. If determining feasibility required computationally expensive simulations, the cost of exploration would be prohibitive. Bayesian search is data-efficient for such problems: starting from a small dataset, the central concept is to use Bayesian models of constraints with an acquisition function to locate promising solutions that may improve predictions of feasibility when the dataset is augmented. At the end of this sequential active learning approach with a limited number of expensive evaluations, the models can accurately predict the feasibility of any solution obviating the need for full simulations. In this paper, we propose a novel acquisition function that combines the probability that a solution lies at the boundary between feasible and infeasible spaces (representing exploitation) and the entropy in predictions (representing explora...

$Research paper thumbnail of A Review of Shape Distortion Methods Available in the OpenFOAM $$^{\textregistered }$$ Framework for Automated Design Optimisation$

A Review of Shape Distortion Methods Available in the OpenFOAM $$^{\textregistered }$$ Framework for Automated Design Optimisation

OpenFOAM®, 2019

Parametrisation of the geometry is one of the essential requirements in shape optimisation, and i... more Parametrisation of the geometry is one of the essential requirements in shape optimisation, and is a challenging subject when carrying out a automated procedure. It is critically important to maintain the consistency of the shape and grid quality between each evaluation, while providing flexibility for a wide range of shapes using the same parameterisation of the geometry. The sensitivity of the grid to the changes to the geometry must be at a minimum during this process. This contribution presents a review of the grid distortion and regeneration methods available within the OpenFOAM\(^{\textregistered }\) framework which can be utilised for shape optimisation. The objective of this contribution is to compare the effectiveness of these methods in the automated procedure and to provide suggestions for improvements. Special attention is given to three major factors involving shape optimisation: automation of model abstraction, automation of grid deformation or regeneration and robustn...

Hybrid Evolutionary Routing Optimisation for Wireless Sensor Mesh Networks

Part of this work was supported by a knowledge transfer partnership (KTP) awarded to the IMC Grou... more

Automatic shape optimisation of the turbine-99 draft tube

This work was supported by the UK Engineering and Physical Sciences Research Council (EPSRC) gran... more

A Bayesian Approach for the Robust Optimisation of Expensive-To-Evaluate Functions

ArXiv, 2019

Many expensive black-box optimisation problems are sensitive to their inputs. In these problems i... more Many expensive black-box optimisation problems are sensitive to their inputs. In these problems it makes more sense to locate a region of good designs, than a single, possible fragile, optimal design. Expensive black-box functions can be optimised effectively with Bayesian optimisation, where a Gaussian process is a popular choice as a prior over the expensive function. We propose a method for robust optimisation using Bayesian optimisation to find a region of design space in which the expensive function's performance is insensitive to the inputs whilst retaining a good quality. This is achieved by sampling realisations from a Gaussian process modelling the expensive function and evaluating the improvement for each realisation. The expectation of these improvements can be optimised cheaply with an evolutionary algorithm to determine the next location at which to evaluate the expensive function. We describe an efficient process to locate the optimum expected improvement. We show ...

Application of multi-objective Bayesian shape optimisation to a sharp-heeled Kaplan draft tube

Optimization and Engineering, 2021

The draft tube of a hydraulic turbine plays an important role for the efficiency and power charac... more The draft tube of a hydraulic turbine plays an important role for the efficiency and power characteristics of the overall system. The shape of the draft tube affects its performance, resulting in an increasing need for data-driven optimisation for its design. In this paper, shape optimisation of an elbow-type draft tube is undertaken, combining Computational Fluid Dynamics and a multi-objective Bayesian methodology. The chosen design objectives were to maximise pressure recovery, and minimise wall-frictional losses along the geometry. The design variables were chosen to explore potential new designs, using a series of subdivision-curves and splines on the inflow cone, outer-heel, and diffuser. The optimisation run was performed under part-load for the Kaplan turbine. The design with the lowest energy-loss identified on the Pareto-front was found to have a straight tapered diffuser, chamfered heel, and a convex inflow cone. Analysis of the performance quantities showed the typically ...

What Makes an Effective Scalarising Function for Multi-Objective Bayesian Optimisation?

Performing multi-objective Bayesian optimisation by scalarising the objectives avoids computation... more Performing multi-objective Bayesian optimisation by scalarising the objectives avoids computation of expensive multi-dimensional integral-based acquisition functions, instead allowing one-dimensional standard acquisition functions—such as Expected Improvement—to be applied. Here, two infill criteria based on hypervolume improvement—one recently-introduced and one novel—are compared with the multi-surrogate Expected Hypervolume Improvement. The reasons for the disparities in these methods’ effectiveness in maximising the hypervolume of the acquired Pareto Front are investigated. In addition, the effect of the surrogate model mean function on exploration and exploitation is examined: careful choice of data normalisation is shown to be preferable to the exploration parameter commonly used with the Expected Improvement acquisition function. Finally, the effectiveness of all the methodological improvements defined here are demonstrated on a real-world problem: the optimisation of a wind ...

Evaluating Mixed-Initiative Procedural Level Design Tools using a Triple-Blind Mixed-Method User Study

Results from a triple-blind mixed-method user study into the effectiveness of mixed-initiative to... more Results from a triple-blind mixed-method user study into the effectiveness of mixed-initiative tools for the procedural generation of game levels are presented. A tool which generates levels using interactive evolutionary optimisation was designed for this study which (a) is focused on supporting the designer to explore the design space and (b) only requires the designer to interact with it by designing levels. The tool identifies level design patterns in an initial hand-designed map and uses that information to drive an interactive optimisation algorithm. A rigorous user study was designed which compared the experiences of designers using the mixed-initiative tool to designers who were given a tool which provided completely random level suggestions. The designers using the mixed-initiative tool showed an increased engagement in the level design task, reporting that it was effective in inspiring new ideas and design directions. This provides significant evidence that procedural cont...

Mixed-Initiative Procedural Content Generation using Level Design Patterns and Interactive Evolutionary Optimisation

ArXiv, 2020

An approach for building mixed-initiative tools for the procedural generation of game levels usin... more An approach for building mixed-initiative tools for the procedural generation of game levels using interactive evolutionary optimisation is introduced. A tool is created based on this approach which (a) is focused on supporting the designer to explore the design space and (b) only requires the designer to interact with it by designing levels. The tool identifies level design patterns in an initial hand-designed map and uses that information to drive an optimisation algorithm. This results in a number of suggestions which are presented to the designer, who can then edit them providing the system with valuable designer feedback. The effectiveness of this approach to create levels with similar level design patterns to a target is illustrated through a series of algorithm driven benchmark tests. To test the mixed-initiative aspect of the tool a triple-blind mixed-method, user study was conducted. When compared to a control group, provided with random level suggestions throughout the des...

Shape optimisation using Computational Fluid Dynamics and Evolutionary Algorithms

This work was supported by the UK Engineering and Physical Sciences Research Council (EPSRC) gran... more

Proceedings of the Genetic and Evolutionary Computation Conference

Optimisation problems often have multiple conflicting objectives that can be computationally and/... more Optimisation problems often have multiple conflicting objectives that can be computationally and/or financially expensive. Monosurrogate Bayesian optimisation (BO) is a popular model-based approach for optimising such black-box functions. It combines objective values via scalarisation and builds a Gaussian process (GP) surrogate of the scalarised values. The location which maximises a cheap-to-query acquisition function is chosen as the next location to expensively evaluate. While BO is an effective strategy, the use of GPs is limiting. Their performance decreases as the problem input dimensionality increases, and their computational complexity scales cubically with the amount of data. To address these limitations, we extend previous work on BO by density-ratio estimation (BORE) to the multi-objective setting. BORE links the computation of the probability of improvement acquisition function to that of probabilistic classification. This enables the use of state-of-the-art classifiers in a BO-like framework. In this work we present MBORE: multi-objective Bayesian optimisation by densityratio estimation, and compare it to BO across a range of synthetic and real-world benchmarks. We find that MBORE performs as well as or better than BO on a wide variety of problems, and that it outperforms BO on high-dimensional and real-world problems. CCS CONCEPTS • Computing methodologies → Modeling methodologies; • Theory of computation → Gaussian processes; • Applied computing → Multi-criterion optimization and decision-making.

On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints

Lecture Notes in Computer Science, Jul 19, 2020

We are often interested in identifying the feasible subset of a decision space under multiple con... more We are often interested in identifying the feasible subset of a decision space under multiple constraints to permit effective design exploration. If determining feasibility required computationally expensive simulations, the cost of exploration would be prohibitive. Bayesian search is data-efficient for such problems: starting from a small dataset, the central concept is to use Bayesian models of constraints with an acquisition function to locate promising solutions that may improve predictions of feasibility when the dataset is augmented. At the end of this sequential active learning approach with a limited number of expensive evaluations, the models can accurately predict the feasibility of any solution obviating the need for full simulations. In this paper, we propose a novel acquisition function that combines the probability that a solution lies at the boundary between feasible and infeasible spaces (representing exploitation) and the entropy in predictions (representing exploration). Experiments confirmed the efficacy of the proposed function.

$Research paper thumbnail of $\epsilon$-shotgun: $\epsilon$-greedy Batch Bayesian Optimisation$

$\epsilon$-shotgun: $\epsilon$-greedy Batch Bayesian Optimisation

arXiv: Learning, Feb 5, 2020

Bayesian Search for Robust Optima

Many expensive black-box optimisation problems are sensitive to their inputs. In these problems i... more Many expensive black-box optimisation problems are sensitive to their inputs. In these problems it makes more sense to locate a region of good designs, than a single—possibly fragile—optimal design. Expensive black-box functions can be optimised effectively with Bayesian optimisation, where a Gaussian process is a popular choice as a prior over the expensive function. We propose a method for robust optimisation using Bayesian optimisation to find a region of design space in which the expensive function’s performance is relatively insensitive to the inputs whilst retaining a good quality. This is achieved by sampling realisations from a Gaussian process that is modelling the expensive function, and evaluating the improvement for each realisation. The expectation of these improvements can be optimised cheaply with an evolutionary algorithm to determine the next location at which to evaluate the expensive function. We describe an efficient process to locate the optimum expected improve...

seanwalton/mixed-initiative-procedural-dungeon-designer: First Release

A unity3d tool which designs dungeon maps based on a level designer's preferences

Shape optimisation of the sharp-heeled Kaplan draft tube: Performance evaluation using Computational Fluid Dynamics

Renewable Energy, 2020

In the first of this two-part contribution, a methodology to assess the performance of an elbow-t... more In the first of this two-part contribution, a methodology to assess the performance of an elbow-type draft tube is outlined. using Computational Fluid Dynamics (CFD) to evaluate the pressure recovery and mechanical energy losses along a draft tube design, while using open-source and commercial software to parameterise and regenerate the geometry and CFD grid. An initial validation study of the elbow-type draft tube is carried out, focusing on the grid-regeneration methodology, and the use of a steady-state assumption for evaluating the design's efficiency. The Grid Convergence Index (GCI) technique was used to assess the uncertainty of the pressure recovery to the grid resolution. It was found that estimating the pressure recovery through area-weighted averaging significantly reduced the uncertainty due to the grid. Simultaneously, it was found that this uncertainty fluctuated with the local cross-sectional area along the geometry. Subsequently, a study of the inflow cone and outer-heel designs on the flowfield and pressure recovery was carried out. Catmull-Rom splines were used to parameterise these components, so as to recreate a number of proposed designs from the literature. GCI analysis is also applied to these designs, demonstrating the robustness of the grid-regeneration methodology.

An interactive tool for enhancing hospital capacity predictions using an epidemiological model

Proceedings of the Genetic and Evolutionary Computation Conference Companion, 2021

Hospital managers have limited resources, and they have to plan on how best to allocate these res... more Hospital managers have limited resources, and they have to plan on how best to allocate these resources based on predicted demand, which is often based on linear or exponential models. In this paper, we propose an interactive tool that produces a forecast of bed occupancy based on an epidemiological model. We optimise this model to fit recently observed data, and interactivity is conferred through a controllable parameter of the model such that users can readily investigate hypothetical scenarios for planning purposes. This study was designed with the Welsh National Health Service, and was born out of their practical need of accurately modelling hospital occupancy during the ongoing Covid-19 pandemic.

Surrogate-Assisted Evolutionary Optimisation (SAEOpt'16) Chairs' Welcome & Organization

Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion - GECCO '16 Companion, 2016

It is our great pleasure to welcome you to the Workshop on Surrogate-Assisted Evolutionary Optimi... more It is our great pleasure to welcome you to the Workshop on Surrogate-Assisted Evolutionary Optimisation (SAEOpt). In many real world optimisation problems evaluating the objective function(s) is computationally expensive. Surrogate-assisted optimisation attempts to alleviate this problem by employing computationally cheap 'surrogate' models to estimate the objective function(s) or the ranking relationships between candidate solutions. Surrogate-assisted approaches have been widely used across the field of evolutionary optimisation, and successful applications include aerodynamic design optimisation, structural design optimisation, data-driven optimisation, chip design, drug design, robotics and many more. Despite recent successes in using surrogate-assisted evolutionary optimisation, there remain many challenges. In this workshop we aim to promote research on surrogate-assisted evolutionary optimisation, particularly the synergies between evolutionary optimisation and machine learning. We hope that the workshop will be of interest to researchers working on the applied and theoretical aspects of surrogate-assisted evolutionary optimisation.

Hybrid Evolutionary Approaches to Maximum Lifetime Routing and Energy Efficiency in Sensor Mesh Networks

Evolutionary Computation, 2015

Mesh network topologies are becoming increasingly popular in battery-powered wireless sensor netw... more Mesh network topologies are becoming increasingly popular in battery-powered wireless sensor networks, primarily because of the extension of network range. However, multihop mesh networks suffer from higher energy costs, and the routing strategy employed directly affects the lifetime of nodes with limited energy resources. Hence when planning routes there are trade-offs to be considered between individual and system-wide battery lifetimes. We present a multiobjective routing optimisation approach using hybrid evolutionary algorithms to approximate the optimal trade-off between the minimum lifetime and the average lifetime of nodes in the network. In order to accomplish this combinatorial optimisation rapidly, our approach prunes the search space using k-shortest path pruning and a graph reduction method that finds candidate routes promoting long minimum lifetimes. When arbitrarily many routes from a node to the base station are permitted, optimal routes may be found as the solution ...

Multi-objective routing optimisation for battery-powered wireless sensor mesh networks

Mesh network topologies are becoming increasingly popular in battery powered wireless sensor netw... more Mesh network topologies are becoming increasingly popular in battery powered wireless sensor networks, primarily due to the extension of network range and resilience against routing failures. However, multi-hop mesh networks suffer from higher energy costs, and the routing strategy directly affects the lifetime of nodes with limited energy sources. Hence while planning routes there are trade-offs to be considered between individual and system-wide battery lifetimes. We present a novel multi-objective routing optimisation approach using evolutionary algorithms to approximate the optimal trade-off between minimum lifetime and the average lifetime of nodes in the network. In order to accomplish this combinatorial optimisation rapidly and thus permit dynamic optimisation for self-healing networks, our approach uses novel k-shortest paths based search space pruning in conjunction with a new edge metric, which associates the energy cost at a pair of nodes with the link between them. We demonstrate our solution on a real network, deployed in the Victoria & Albert Museum, London. We show that this approach provides better trade-off solutions in comparison to the minimum energy option, and how a combination of solutions over the lifetime of the network can enhance the overall minimum lifetime.

On Bayesian Search for the Feasible Space Under Computationally Expensive Constraints

We are often interested in identifying the feasible subset of a decision space under multiple con... more We are often interested in identifying the feasible subset of a decision space under multiple constraints to permit effective design exploration. If determining feasibility required computationally expensive simulations, the cost of exploration would be prohibitive. Bayesian search is data-efficient for such problems: starting from a small dataset, the central concept is to use Bayesian models of constraints with an acquisition function to locate promising solutions that may improve predictions of feasibility when the dataset is augmented. At the end of this sequential active learning approach with a limited number of expensive evaluations, the models can accurately predict the feasibility of any solution obviating the need for full simulations. In this paper, we propose a novel acquisition function that combines the probability that a solution lies at the boundary between feasible and infeasible spaces (representing exploitation) and the entropy in predictions (representing explora...

$Research paper thumbnail of A Review of Shape Distortion Methods Available in the OpenFOAM $$^{\textregistered }$$ Framework for Automated Design Optimisation$

A Review of Shape Distortion Methods Available in the OpenFOAM $$^{\textregistered }$$ Framework for Automated Design Optimisation

OpenFOAM®, 2019

Parametrisation of the geometry is one of the essential requirements in shape optimisation, and i... more Parametrisation of the geometry is one of the essential requirements in shape optimisation, and is a challenging subject when carrying out a automated procedure. It is critically important to maintain the consistency of the shape and grid quality between each evaluation, while providing flexibility for a wide range of shapes using the same parameterisation of the geometry. The sensitivity of the grid to the changes to the geometry must be at a minimum during this process. This contribution presents a review of the grid distortion and regeneration methods available within the OpenFOAM\(^{\textregistered }\) framework which can be utilised for shape optimisation. The objective of this contribution is to compare the effectiveness of these methods in the automated procedure and to provide suggestions for improvements. Special attention is given to three major factors involving shape optimisation: automation of model abstraction, automation of grid deformation or regeneration and robustn...

Hybrid Evolutionary Routing Optimisation for Wireless Sensor Mesh Networks

Part of this work was supported by a knowledge transfer partnership (KTP) awarded to the IMC Grou... more

Automatic shape optimisation of the turbine-99 draft tube

This work was supported by the UK Engineering and Physical Sciences Research Council (EPSRC) gran... more

A Bayesian Approach for the Robust Optimisation of Expensive-To-Evaluate Functions

ArXiv, 2019

Many expensive black-box optimisation problems are sensitive to their inputs. In these problems i... more Many expensive black-box optimisation problems are sensitive to their inputs. In these problems it makes more sense to locate a region of good designs, than a single, possible fragile, optimal design. Expensive black-box functions can be optimised effectively with Bayesian optimisation, where a Gaussian process is a popular choice as a prior over the expensive function. We propose a method for robust optimisation using Bayesian optimisation to find a region of design space in which the expensive function's performance is insensitive to the inputs whilst retaining a good quality. This is achieved by sampling realisations from a Gaussian process modelling the expensive function and evaluating the improvement for each realisation. The expectation of these improvements can be optimised cheaply with an evolutionary algorithm to determine the next location at which to evaluate the expensive function. We describe an efficient process to locate the optimum expected improvement. We show ...

Application of multi-objective Bayesian shape optimisation to a sharp-heeled Kaplan draft tube

Optimization and Engineering, 2021

The draft tube of a hydraulic turbine plays an important role for the efficiency and power charac... more The draft tube of a hydraulic turbine plays an important role for the efficiency and power characteristics of the overall system. The shape of the draft tube affects its performance, resulting in an increasing need for data-driven optimisation for its design. In this paper, shape optimisation of an elbow-type draft tube is undertaken, combining Computational Fluid Dynamics and a multi-objective Bayesian methodology. The chosen design objectives were to maximise pressure recovery, and minimise wall-frictional losses along the geometry. The design variables were chosen to explore potential new designs, using a series of subdivision-curves and splines on the inflow cone, outer-heel, and diffuser. The optimisation run was performed under part-load for the Kaplan turbine. The design with the lowest energy-loss identified on the Pareto-front was found to have a straight tapered diffuser, chamfered heel, and a convex inflow cone. Analysis of the performance quantities showed the typically ...

What Makes an Effective Scalarising Function for Multi-Objective Bayesian Optimisation?

Performing multi-objective Bayesian optimisation by scalarising the objectives avoids computation... more Performing multi-objective Bayesian optimisation by scalarising the objectives avoids computation of expensive multi-dimensional integral-based acquisition functions, instead allowing one-dimensional standard acquisition functions—such as Expected Improvement—to be applied. Here, two infill criteria based on hypervolume improvement—one recently-introduced and one novel—are compared with the multi-surrogate Expected Hypervolume Improvement. The reasons for the disparities in these methods’ effectiveness in maximising the hypervolume of the acquired Pareto Front are investigated. In addition, the effect of the surrogate model mean function on exploration and exploitation is examined: careful choice of data normalisation is shown to be preferable to the exploration parameter commonly used with the Expected Improvement acquisition function. Finally, the effectiveness of all the methodological improvements defined here are demonstrated on a real-world problem: the optimisation of a wind ...

Evaluating Mixed-Initiative Procedural Level Design Tools using a Triple-Blind Mixed-Method User Study

Results from a triple-blind mixed-method user study into the effectiveness of mixed-initiative to... more Results from a triple-blind mixed-method user study into the effectiveness of mixed-initiative tools for the procedural generation of game levels are presented. A tool which generates levels using interactive evolutionary optimisation was designed for this study which (a) is focused on supporting the designer to explore the design space and (b) only requires the designer to interact with it by designing levels. The tool identifies level design patterns in an initial hand-designed map and uses that information to drive an interactive optimisation algorithm. A rigorous user study was designed which compared the experiences of designers using the mixed-initiative tool to designers who were given a tool which provided completely random level suggestions. The designers using the mixed-initiative tool showed an increased engagement in the level design task, reporting that it was effective in inspiring new ideas and design directions. This provides significant evidence that procedural cont...

Mixed-Initiative Procedural Content Generation using Level Design Patterns and Interactive Evolutionary Optimisation

ArXiv, 2020

An approach for building mixed-initiative tools for the procedural generation of game levels usin... more An approach for building mixed-initiative tools for the procedural generation of game levels using interactive evolutionary optimisation is introduced. A tool is created based on this approach which (a) is focused on supporting the designer to explore the design space and (b) only requires the designer to interact with it by designing levels. The tool identifies level design patterns in an initial hand-designed map and uses that information to drive an optimisation algorithm. This results in a number of suggestions which are presented to the designer, who can then edit them providing the system with valuable designer feedback. The effectiveness of this approach to create levels with similar level design patterns to a target is illustrated through a series of algorithm driven benchmark tests. To test the mixed-initiative aspect of the tool a triple-blind mixed-method, user study was conducted. When compared to a control group, provided with random level suggestions throughout the des...

Shape optimisation using Computational Fluid Dynamics and Evolutionary Algorithms

This work was supported by the UK Engineering and Physical Sciences Research Council (EPSRC) gran... more