Influence of the Working Strategy on A-Team Performance

An A-Team is a system of autonomous agents and the common memory. Each agent possesses some problem-solving skills and the memory contains a population of problem solutions. Cyclically solutions are being sent from the common memory to agents and from agents back to the common memory. Agents cooperate through selecting and modifying these solutions according to the user-defined strategy referred to as the working strategy. The modifications can be constructive or destructive. An attempt to improve a solution can be successful or unsuccessful. Agents can work asynchronously (each at its own speed) and in parallel. The A-Team working strategy includes a set of rules for agent communication, selection of solution to be improved and management of the population of solutions which are kept in the common memory. In this paper influence of different strategies on A-Team performance is investigated. To implement various strategies the A-Team platform called JABAT has been used. Different working strategies with respect to selecting solutions to be improved by the A-Team members and replacing the solutions stored in the common memory by the improved ones are studied. To evaluate typical working strategies the computational experiment has been carried out using several benchmark data sets. The experiment shows that designing effective working strategy can considerably improve the performance of the A-Team system.

[1]  John Rachlin,et al.  A-Teams: An Agent Architecture for Optimization and Decision Support , 1998, ATAL.

[2]  Azim Eskandarian,et al.  ENHANCING INTELLIGENT AGENT COLLABORATION FOR FLOW OPTIMIZATION OF RAILROAD TRAFFIC , 2002 .

[3]  Piotr Jedrzejowicz,et al.  Web Accessible A-Team Middleware , 2008, ICCS.

[4]  Brian W. Kernighan,et al.  An Effective Heuristic Algorithm for the Traveling-Salesman Problem , 1973, Oper. Res..

[5]  Pedro S. de Souza,et al.  Asynchronous Teams: Cooperation Schemes for Autonomous Agents , 1998, J. Heuristics.

[6]  Enrique H. Ruspini,et al.  Numerical methods for fuzzy clustering , 1970, Inf. Sci..

[7]  Jaime Simão Sichman,et al.  Using A-Teams to optimize automatic insertion of electronic components , 2003, Adv. Eng. Informatics.

[8]  Panos M. Pardalos,et al.  A parallel implementation of an asynchronous team to the point-to-point connection problem , 2003, Parallel Comput..

[9]  Piotr Jedrzejowicz,et al.  JADE-Based A-Team as a Tool for Implementing Population-Based Algorithms , 2006, Sixth International Conference on Intelligent Systems Design and Applications.

[10]  Nicholas R. Jennings,et al.  Agent Theories, Architectures, and Languages: A Survey , 1995, ECAI Workshop on Agent Theories, Architectures, and Languages.

[11]  Piotr Jędrzejowicz,et al.  Population Learning Algorithm for the Resource-Constrained Project Scheduling , 2006 .

[12]  Piotr Jedrzejowicz,et al.  JADE-Based A-Team Environment , 2006, International Conference on Computational Science.

[13]  PAUL C. BADAVAS,et al.  Response identification of distributed systems with noisy measurements at finite points , 1970, Inf. Sci..

[14]  Stephen Gilmore,et al.  Combining Measurement and Stochastic Modelling to Enhance Scheduling Decisions for a Parallel Mean Value Analysis Algorithm , 2006, International Conference on Computational Science.

[15]  Marco Aldinucci,et al.  Computational Science - ICCS 2008, 8th International Conference, Kraków, Poland, June 23-25, 2008, Proceedings, Part I , 2008, ICCS.

[16]  William H. K. Lam,et al.  A capacity restraint transit assignment with elastic line frequency , 2002 .

[17]  Ibrahim H. Osman,et al.  Metastrategy simulated annealing and tabu search algorithms for the vehicle routing problem , 1993, Ann. Oper. Res..

[18]  Roman Neruda,et al.  Bang 3: a computational multi-agent system , 2004 .