作业调度问题的合作求解方法研究

A study on cooperation-solving approach to solving the job-scheduling problem

作业调度问题是一类典型的困难组合优化问题 ,基于知识的多智能主体系统为解决此类问题提供了一种新的思路 .在生产作业管理中引入多智能主体技术 ,可以很方便地用智能主体及多智能主体系统描述制造系统及其组成要素的行为与运行机制 ,作业调度通过智能主体之间及多智能主体系统的合作加以实现 ,达到预定的运行目标和状态 .构造了单个智能主体的结构 ,对作业调度问题中存在的三种主要冲突类型与基于知识的协商消除冲突的方法及其步骤进行了探讨 .

The job scheduling problem (JSP) is a typically hard nonpolynomial problem (NP hard), and the multi agent system (MAS) based on knowledge is a new solution to JSP. The behavior and operating mechanism of the units of manufacturing system would be conveniently described and simulated by using agent and MAS. The job scheduling would be accomplished by agent cooperating with each other and the cooperation of MAS , and the due operating goals and status would be achieved. The framework of the single agent is constructed. The three kinds of conflicts existing in JSP are discussed and the negotiation approach and process are studied based on knowledge to remove conflicts. The Multi agent Cooperation Solving Approach (MCSA) based on knowledge is created.