摘要
针对虚拟企业环境下生产调度具有动态变化的特点,采用多智能体技术,构建了虚拟企业任务调度运行模型。模型中引入了一种动态任务分配技术,以资源智能体所承担的生产任务为对象,构建了以制造时间最小化为目标的Markov博弈模型,并用一系列静态势力场博弈来近似Markov博弈,从而减少了算法的复杂性。实验表明,该调度方法基本等同于集中式任务分配法,并且具有更佳的通信和交流鲁棒性。
Aiming at the feature of dynamic variation of the production scheduling under virtual enterprise environment,by adopting multi-agent technology,the task scheduling and operation model for virtual enterprise is built.In the model the dynamic task allocation technology has been introduced,with the production task of the resource agent as the object,the Markov game model is established upon the minimized manufacturing time as the target.In addition,by using series of static potential games for approximating Markov game to decrease the complexity of the algorithm.The experiment verifies that this method is basically equivalent to the centralized task allocation method and features better robustness of communication and exchange.
出处
《自动化仪表》
CAS
北大核心
2011年第1期39-41,共3页
Process Automation Instrumentation
关键词
任务调度
虚拟企业
多智能体
通信
调度仿真
Task scheduling Virtual enterprise(VE) Multi-agent Communication Scheduling simulation