ABDIDSS环境下基于多Agents协作的任务求解

Task solving based on multi-agents cooperation in the environment of ABDIDSS

传统的DSS不能满足复杂的非结构化决策问题的要求,ABDIDSS是由若干个不同的“Agent-Region”的ABIDSSes通过网络连接而成,根据ABDIDSS中“Agent-Region”具有的知识,把多Agents之间的协作分成任务分担的协作与结果共享的协作;前者采用合同网协议实现任务、子任务在多Agents之间的求解,后者采用局部全局规划模式实现任务协作求解;文章阐述了在ABDIDSS环境中决策任务求解过程涉及的单个Agent行为控制、“Agent-Region”内Agents合作求解子任务以及分布式环境下不同“Agent-Region”的Agents协作求解整体任务的过程与方法。

As the classical decision support system（DSS） can not satisfy the requirement of complex and non-structuring decision problems ,the agent-based distributed intelligent decision support system （ABDIDSS） is put forward which consists of some ABIDSSes in the various “Agent-Regions” connected through Internet. According to the knowledge provided by the“Agent-Region” in the ABDIDSS, the cooperation among multi-agents is divided into task-sharing cooperation and results-sharing cooperation while the former uses contract net protocol to realize the solving of tasks and sub-tasks among multi-agents and the latter uses partial global planning to realize cooperative solving of tasks among multi-agents. The processes and methods concerned with individual agent behavior control ,cooperative solving of sub-tasks by agents in “Agent-Region” and solving of global tasks by agents in various “Agent-Regions” in the distributed environment are put forward in the environment of the ABDIDSS.