基于多Agent和遗传算法的制造网格资源调度  

Manufacturing grid scheduling based on multi-agent and genetic algorithm

作　　者：赵付青[1] 李瑞生[1] 

机构地区：[1]兰州理工大学计算机与通信学院,甘肃兰州730050

出　　处：《兰州理工大学学报》2007年第6期85-88,共4页Journal of Lanzhou University of Technology

基　　金：甘肃省自然科学基金(3ZS062-B25-033)

摘　　要：为了保障用户的多QoS要求,在制造网格资源调度中引入灵活健壮的协商机制和高效的资源调度算法尤显重要.为此提出基于多Agent的制造网格资源调度方法,运用合同网协议作为多Agent的协作协议并根据制造网格资源调度需求运用遗传算法作为资源优化组合的算法.综合应用多Agent开发工具JADE和遗传算法工具包JGAP进行模拟调度实验,证明了方法的有效性.In order to satisfy the multi-QoS requirement of the users in the manufacturing grid scheduling, a flexible and robust negotiation mechanism and scheduling algorithm were needed. Therefore a multiagent based scheduling approach for the manufacturing grid was proposed. Contract net protocol for multiagent cooperation and negotiation mechanism were employed, and then a genetic algorithm as scheduling algorithm was also used. An experiment was conducted based on JADE and JGAP to testify this approach. Simulation result showed that the method was effective.

关 键 词：制造网格 多AGENT 资源调度 遗传算法 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]