论文部分内容阅读
Agent联盟形成是多Agent系统的一个基本操作.多个Agent往往需要形成联盟来协作完成任务或者提高联盟体的能力.现有的Agent联盟算法并不考虑任务内部的结构.本文则考虑一类具有复杂内部结构的任务.任务由一组子任务构成,子任务之间存在着逻辑依赖关系,并且以BPMN的流程形式描述.本文还在此基础上加入了对相邻子任务之间的转移成本的考虑.并且本文提出了一种基于动态规划的多项式时间算法来解决此类加入了逻辑依赖关系和转移成本因素的Agent联盟问题.
Agent alliance formation is a basic operation of multi-Agent system.Many Agents often need to form alliances to collaborate to accomplish the task or improve the alliance’s ability.The existing Agent alliance algorithm does not consider the internal structure of the task.In this paper, The task of complex internal structure is composed of a group of subtasks, and there is a logical dependency between subtasks and is described in the form of BPMN flow.This paper also adds the transition cost between adjacent subtasks This paper proposes a polynomial-time algorithm based on dynamic programming to solve the problem of Agent alliance with such factors as logical dependency and transfer cost.