Journal of South China University of Technology (Natural Science Edition) ›› 2008, Vol. 36 ›› Issue (9): 11-14,30.

• Computer Science & Technology • Previous Articles     Next Articles

Fast Generation Algorithm of Agent Coalition in Parallel Multi-Task Environment

Hao Zhi-feng  Cai Rui-chu   

  1. School of Computer Science and Engineering, South China University of Technology, Guangzhou 510640, Guangdong, China
  • Received:2007-10-15 Revised:2008-06-05 Online:2008-09-25 Published:2008-09-25
  • Contact: 郝志峰（1968-），男，教授，博士生导师，主要从事机器学习、人工智能等研究． E-mail:mazfhao@scut．edu．cn
  • About author:郝志峰（1968-），男，教授，博士生导师，主要从事机器学习、人工智能等研究．
  • Supported by:

    教育部新世纪优秀人才基金资助项目（NCET-05-0734）;广东省自然科学基金资助项目（04020079）;南京大学软件国家重点实验室开放基金资助项目（200603）;东南大学移动通信国家重点实验室开放基金资助项目（A200605）;广东省科技攻关项目（2005B10101010）

PDF

1420

Abstract:

A multi-colony ant colony optimization （MCACO） is proposed for the coalition generation problem in the parallel multi-task environment. In this algorithm, the ants from the same colony cooperate to accomplish a task by reasonably allocating the resources, while those from different colonies scramble the resources and resolve the resource conflict among multiple parallel tasks. Moreover, a specially-designed pheromone update rule is applied to enhance the global search ability of MCACO and improve the generation quality of agent coalition by taking the tradeoff between the local optimization benefit and the global one. Simulation results indicate that the proposed rithm helps to generate more effective coalition in multiple benchmark environments.

Key words: muhi-Agent system, multiple task, coalition generation, ant colony algorithm