Distributed Constraint Satisfaction For Coordinating And Integrating A Large Scale Heterogeneous Enterprise

Distributed Constraint Satisfaction for Coordinating and Integrating a Large-scale, Heterogeneous Enterprise

Artificial Intelligence: Methodology, Systems, and Applications

This book constitutes the refereed proceedings of the 13th International Conference on Artificial Intelligence: Methodology, Systems, and Applications, AIMSA 2008, held in Varna, Bulgaria in September 2008. The 30 revised full papers presented together with the 10 posters were carefully reviewed and selected from 109 submissions. The papers are organized in topical sections on agents; natural language processing and text analysis; machine learning and information retrieval; knowledge representation and reasoning; constraints, heuristics and search; applications; posters.

Research and Development in Intelligent Systems XXIV

An agent in a multi-agent system (MAS) has to generate plans for its individual goal, but these plans may con?ict with those that are already being scheduled or executed by other agents. It must also be able to complete its planning and resolution of these con?icts within a reasonable time to have an acceptable quality plan. Although we adopt hierarchical planning (HP, for example, see [7, 12]) using the decision-theoretic planning (DTP) approach [6] for ef?cient planning, it is not trivial to apply HPO to MAS. In HP, appropriate (abstract) plans are selected level by level to maximize the utility U (p), where where p is the expected ?nal plan comprising a sequence of primitive actions. However, in the MAS context, con?icts between agents affect the ef?ciency and quality of resulting plans. When a con?ict is found at lower levels, an additional sophisticated process for avoiding it (con?ict resolution) must be invoked and some extra actions (such as waiting for synchronization and detouring) may have to be added to the plan. The con?ict resolution process may become costly or fail. Even a single con?ict, if it is dif?cult to resolve, will result in a plan with considerably lower quality than it otherwise would have. As a result, in multi-agent systems, the second- or third-best plans may result in better overall performance.