集成预防性维护和流水线调度的鲁棒性优化研究

Integrating Production Scheduling and Maintenance Policy for Robustness in Flow Shop Problems

针对离散流水车间,设备故障率函数服从威布尔分布,在考虑维护策略的基础上,以工件的最终完工时间期望值为质量鲁棒性指标、以所有工序的开始加工时间的延迟总和的期望值为解鲁棒性指标,建立了不确定性环境下预防性维护(Preventive maintenance,PM)和生产调度的集成优化模型,联合决策各工序的开始加工时间和预防性维护位置.进一步,设计了基于工件优先列表、有效代理指标、邻域搜索机制的三阶段启发式算法对模型进行求解.最后,数值实验与传统方法对比结果表明,系统最优缓冲时间随着解鲁棒性权重的增大而逐渐增加,且质量鲁棒性堕化速度远小于解鲁棒性提升的速度,使得其与传统方法相比总体目标愈加优异.

For the flow-shops, where the machines0 failure function is governed by the Weibull distribution, considering the maintenance strategy, a joint model of integrating run-based preventive maintenance （PM） and production scheduling is proposed under the uncertainty environment, in which the planned start times of jobs and the PM times are determined simultaneously. And, the makespan is selected as the quality robustness measure;the total delay of the jobs0 start time is selected as the solution robustness measure. Then, a three-phase heuristic algorithm based on the priority list, surrogate measure, and local search is devised to solve the mathematic model. Experimental results demonstrate that the solution robustness can be significantly improved at the cost of very little degradation in quality robustness using our algorithm compared with the traditional way.