不确定环境下的多Agent鲁棒性预测调度研究

Predictable Robust Scheduling under Uncertainty-a Multi-Agent Based Approach

由于生产过程中的不确定性事件及其影响都具有随机性,因此难以建立精确数学模型以实现调度问题的求解。针对不确定环境下的Job Shop调度,分析生产干扰及其扩散效应的随机性特征,根据不确定性事件集与设备运行效率之间的内在关系,提出了一种基于设备整体效能(OEE)的具有鲁棒性的预测调度实现方法,并设计了一种基于多Agent协作机制的预测调度系统模型。仿真试验表明,该模型可以减少工序延误,提高调度方案的稳定性。

Uncertainty events and their influences in a real manufacturing situation are a complex stochastic phenomenon, and it is very difficult to build accurate formulations used in addressing the scheduling problem and in solution processes. Some issues related to executing production schedules in the presence of unforeseen disruptions in the job shop were discussed, the interrelationship among uncertainties and equipment effectiveness was deduced, and a novel robust approach to solve the predictable scheduling problem based on overall equipment effectiveness（OEE） was presented. An agent--based framework used for predictable scheduling system was also proposed. Finally, the simulation results illustrate that the proposed approach is effective in reducing the delay of operation and in enhancing the stability of scheduling.