基于离散粒子群求解复杂联盟的并行生成ComplicatedCoalition.pdfVIP

第 31 卷第 3 期 电 子 与 信 息 学 报 Vol.31No.3 2009 年 3 月 Journal of Electronics Information Technology Mar.2009 基于离散粒子群求解复杂联盟的并行生成 蒋建国 张国富 齐美彬 苏兆品 (合肥工业大学计算机与信息学院 合肥 230009) (安全关键工业测控技术教育部工程研究中心 合肥 230009) 摘 要：联盟生成是多 agent 系统中的一个关键问题。该文引入离散粒子群优化来解决这一问题，采用粒子的随机 扰动避免了算法的早熟，设计一种二维二进制编码实现复杂联盟的并行生成，通过编码可行性检查、冲突消解和补 偿策略克服了求解过程中的资源冲突和联盟死锁。仿真实验说明了算法的有效性。 关键词：复杂联盟；离散粒子群优化；并行生成；冲突消解 中图分类号：TP18 文献标识码： A 文章编号：1009-5896(2009)03-0519-04 Complicated Coalition Parallel Generation Based on Discrete Particle Swarm Optimization Jiang Jian-guo Zhang Guo-fu Qi Mei-bin Su Zhao-pin (School of Computer and Information Science, Hefei University of Technology, Hefei 230009, China) (Engineering Research Center of Safety Critical Industrial Measurement and Control Technology, Ministry of Education, Hefei 230009, China) Abstract: Coalition generation is a key problem in multi-agent systems. Discrete particle swarm optimization is adopted to solve the problem, and random perturbation of particles is used to avoid premature convergence. A novel two-dimensional binary encoding is designed to realize complicated coalition parallel generation, and the operation of conflict resolution, feasibility checking, and compensation strategy are brought into effect to surmount the resource conflict and coalition lock in the course of problem solving. Finally, several typical simulation studies are carried out to illustrate its validity. Key words: Complicated coalition; Discrete particle swarm optimization; Parallel generation; Conflict resolution