Agreeing on Defeasible Commitments

Advances in Agent Communication, 2004

In this paper we address the problem of how the autonomy of agents in an organization can be enhanced by means of contracts. Contracts are modelled as legal institutions: systems of legal rules which allow to change the regulative and constitutive rules of an organization. The methodology we use is to attribute to organizations mental attitudes, beliefs, desires and goals, and to take into account their behavior by using recursive modelling.

2005

For over a decade, agent research has shown that social commitments support the definition of open multiagent systems by capturing the responsibilities that agents contract toward one another through their communications. These systems, however, rely on the assumption that agents respect the social commitments they adopt. To overcome this limitation, in this paper we investigate the role of sanctions as elements whose enforcement fosters agents' compliance with adopted commitments.

Applied Intelligence, 2013

Commitments are being used to specify interactions among autonomous agents in multiagent systems. Various formalizations of commitments have shown their strength in representing and reasoning on multiagent interactions. These formalizations mostly study commitment lifecycles, emphasizing fulfillment of a single commitment. However, when multiple commitments coexist, fulfillment of one commitment may have an effect on the lifecycle of other commitments. Since agents generally participate in more than one commitment at a time, it is important for an agent to determine whether it can honor its commitments. These commitments may be the existing commitments of the agent as well as any prospective commitments that the agent plans to participate in. To address this, we develop the concept of commitment feasibility, i.e., whether it is possible for an agent to fulfill a set of commitments all together. To achieve this we generalize the fulfillment of a single commitment to the feasibility of a set of commitments. We then develop a solid method to determine commitment feasibility. Our method is based on the transformation of feasibility into a constraint satisfaction problem and use of constraint satisfaction techniques to come up with a conclusion. We show soundness and completeness of our method and illustrate its applicability over realistic cases. Keywords multiagent interaction • commitments • fulfillment • feasibility 1 Introduction Multiagent systems operate through agents' interactions. These interactions are mostly bidirectional, where agents do things for others and expect things in return, mimicking the workings of traditional business. A commitment is an important abstraction to represent

28

2015

Previous work describes the use of formal commitments to mediate the communication between autonomous agents through commitment-based protocols. I extend that work to examine the conditions that encourage the success of agents that use commitments. I define a parameterized and iterated Committer’s Dilemma game that extends the well-known Prisoner’s Dilemma game, and then use this game and different agent strategies to examine how the conditions for commitments affect game outcomes. I describe the results of multiple simulations on a multiagent society with various game parameters. Results show that committing and satisfying agent types dominate over other agent types most frequently (1) when commitments are frequently exchanged, (2) when games tend to end at other than a Nash equilibria, (3) when the cost to create a commitment is low, and (4) when the utility of a given good is about 40% to 60% of the utility of a received good. A classifier, with over 95% accuracy, is trained to p...

Lecture Notes in Computer Science, 2012

This paper presents an action language, called L mt , for representing and reasoning about commitments in multi-agent domains. The language is an extension of the language L, with new features motivated by the problem of representing and reasoning about commitments. These features include time, delayed effects, ir/reversible effects, concurrent actions, and multi-agents, for specifying and reasoning about narratives in multi-agent domains. The paper provides a transition-based semantics for L mt , which makes it possible to define an entailment relation between queries and multi-agent narratives with time constraints. The paper also demonstrates how features and properties of commitments can be described in this action language. In particular, it shows how L mt can handle both simple commitment actions as well as complex commitment protocols. Furthermore, the semantics of L mt provides a uniform solution to different problems in reasoning about commitments such as the problem of (i) verifying whether an agent fails (or succeeds) to deliver on its commitments; (ii) identifying outstanding commitments; and (iii) suggesting ways to satisfy outstanding commitments. Introduction and Motivation Consider the following conversation between agents A and B: Agent A: Do you want to do something tonight? Agent B: Sure, what do you want to do? Agent A: Let us have a potluck dinner with X. I will prepare some sandwiches and call X. But can you pick her up? Also, could you bring some soft-drinks? Agent B: Sure. How about 7pm? Agent A: Great. The conversation highlights a number of activities that A and B promise to perform: A needs to prepare the sandwiches and call X. These activities need to be completed before 7pm. B, on the other hand, needs to show up at A's flat by 7pm with soft-drinks and with X. These activities are referred to as commitments between A and B. This conversation also provides a number of interesting questions. What happens if A fails to make the phone call to X? What happens if B does not have enough money to buy the drinks? Can B ask A for money or can B ask X to bring the softdrinks? More generally, what does it mean for an agent to

Proceedings of the first international joint …, 2002

Applied Artificial Intelligence, 2004

Applied Intelligence, 2014

Both knowledge and social commitments have received considerable attention in Multi-Agent Systems (MASs), specially for multi-agent communication. Plenty of work has been carried out to define their semantics. However, the relationship between social commitments and knowledge has not been investigated yet. In this paper, we aim to explore such a relationship from the semantics and model checking perspectives with respect to CTLK logic (an extension of CTL logic with modality for reasoning about knowledge) and CTLC logic (an extension of CTL with modalities for reasoning about commitments and their fulfillments). To analyze this logical relationship, we simply combine the two logics in one new logic named CTLKC. The purpose of such a combination is not to advocate a new logic, but only to express and figure out some reasoning postulates merging both knowledge and commitments as they are currently defined in the literature. By so doing, we identify some paradoxes in the new logic showing that simply combining current versions of commitment and knowledge

Agreement Technologies, 2012

Research on negotiation and task allocation has been in the multi-agent systems realm since its inception as a research field. More recently, social aspects of agenthood have received increasing attention, namely developing on the fields of normative and trust systems. The integration of these different research contributions will allow to build robust applications for electronic agreement negotiation, aiming at their acceptability and application in industry. The ANTE 1 framework is the corollary of an ongoing long-term research project that encompasses three main agreement technologies: negotiation [3], normative environments [1], and computational trust [5]. Although ANTE has been targeting the domain of B2B electronic contracting, it was conceived as a more general framework having in mind a wider range of applications. This paper describes a demonstration showing the application of the ANTE framework to an agent-based automatic electronic contracting domain. 2 Main purpose ANTE addresses the issue of multi-agent collective work in a comprehensive way, covering both negotiation as a mechanism for finding mutually acceptable agreements, and the enactment of such agreements. It also includes the evaluation of the enactment phase, with the aim of improving future negotiations. This demonstration shows an application scenario where three research areas-negotiation, normative

Languages, Methodologies and Development Tools for Multi-Agent Systems, 2009

Existing approaches about defining formal semantics of commitment usually consider operations as axioms or constrains on top of the commitment semantics, which fail to capture the meaning of interactions that are central to real-life business scenarios. Furthermore, existing semantic frameworks using different logics do not gather the full semantics of commitment operations and semantics of social commitments within the same

Lecture Notes in Computer Science, 2007

Agent Communication Languages (ACLs) play a fundamental role in open multiagent systems where message exchange is the main if not the only way for agents to coordinate themselves. New proposals about ACL semantics based on social commitments aim at countering the shortcomings of the mainstream mental-state-based ones. The commitment solution does not come for free and calls for an adequate monitoring system that checks whether commitments are fulfilled or not.

2013

We relate two contract models: one based on event structures and game theory, and the other one based on logic. In particular, we show that the notions of agreement and winning strategies in the game-theoretic model are related to that of provability in the logical model.

2004

Commitments are a powerful abstraction for representing the interactions between agents. Commitments capture the content of the interactions declaratively and allow agents to reason about their actions. Recent work on multiagent protocols define agent interactions as the creation and manipulation of commitments to one another. As a result, commitment protocols can be executed flexibly, enabling the agents to cope with exceptions that arise at run time.