摘要
Agent联盟构成问题是一个复杂的组合优化问题,属于NP完全问题,其求解是很困难的。本文提出了一种Agent联盟构成策略,该策略在系统运行的初始阶段,保证任务分配的最优解。而在随后的运行中基于熟人集形成Agent联盟。该策略保证在接近任务分配优化的情况下,有效减少系统中的可能联盟数以及联盟形成过程中的通信开销和计算量,避免联盟形成过程中的盲目性,从而节省协商时间提高协商效率。
Agent Coalition Formation is a complicated problem of combinatorial optimization which belongs to NP-Complete problems. It is very difficult to get a solution. We proposed a strategy of coalition formation. This method may produce an optimal solution at the beginning of the system works. And then agents coalition is formed based on the set of acquaintance. The method assures that the system form a coalition close to optimal result at a less communication and calculation cost. It can avoid blindness in coalition forming process and improve the efficiency of negotiation.
出处
《电光与控制》
北大核心
2007年第3期147-149,153,共4页
Electronics Optics & Control
作者简介
程红斌（1978-），男，山西闻喜人，博士生，主要研究领域为信息系统工程与智能决策。
