权重式多Agent作业计划及其在Job Shop中的应用

A Weighted Multi-Agent Scheduling Method and Its Applications in Job Shops

提出了一种新的分布式多Agent通用作业计划编制体系。为体系中每个设备、任务和控制中心设计了计划编制Agent,以计算竞争作业的评价指数 ,根据车间的特征决定各Agent的权重 ,并用加权平均方式确定各竞争作业的最终评价指数。将这种一般策略应用于JobShop生产系统的最小交工延迟问题 ,与以往的启发式算法相比 。

A new general distributed multi-agent production scheduling system was introduced, in which each machine, job and control center had its own scheduling agent. Whenever a scheduling decision is needed, each relevant agent calculates an agent priority for all competing jobs. The system determines each agent’s weight based on system conditions, and obtains final priority as a weighted average of agent priorities. The application of the general scheme in a job shop with minimizing mean tardiness objective shows favorable performance of the method over traditional heuristics.