生产系统干扰应对策略与重调度集成优化方法研究

项目名称： 生产系统干扰应对策略与重调度集成优化方法研究

项目编号： No.71501024

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 管理科学

项目作者： 王杜娟

作者单位： 大连海事大学

项目金额： 15.5万元

中文摘要： 在具有恶化效应和学习效应的生产系统中，针对机器故障、工件延时到达和客户工期调整典型干扰事件的影响，从系统多主体利益出发，辨识资源分配、预防性维护和加工可拒绝等干扰应对策略，并提出扰动度量方法。从干扰应对策略与重调度集成优化的视角，研究干扰管理问题。构建资源分配与重调度集成优化模型、预防性维护与重调度集成优化模型和加工可拒绝与重调度集成优化模型。分析模型特性和Pareto最优解相关的先验知识；提出基于代理模型进行适应度评估的多目标进化算法，并引入先验知识以引导优化过程。通过数值实验和应用检验，验证模型和算法的有效性。本项目不仅将原有针对加工时间固定生产调度的干扰管理研究拓展到加工时间可变的生产环境，而且提出资源分配、预防性维护和加工可拒绝与重调度的集成优化方法。研究成果将在理论上有效解决加工时间可变干扰管理问题，丰富和发展多目标进化算法，同时为加工制造企业应对干扰事件提供理论和方法支持。

中文关键词： 干扰管理；集成优化；多目标进化算法；代理模型；先验知识

英文摘要： In production scheduling with deteriorating and learning effects, disruption reaction strategies, such as resource allocation, preventive maintenance and job rejection etc, are identified considering interests of management, shop-floor operators and customers to address the impact of typical disruption events, namely, machine breakdown, job arrival delay and changes in customer due dates, and corresponding deviation measurements are proposed.Disruption management is investigated from the perspective of integrated optimization of reaction strategies and rescheduling. Integrated optimization models of resource allocation, preventive maintenance and job rejection with rescheduling are constructed. To effectively solve the models, multi-objective evolutionary algorithm is designed and improved by embedding surrogate model for fitness approximation, and by utilizing apriori knowledge about Pareto optimal solution to guide evolution. The effectiveness of models and algorithms is tested through numerical studies and application verification. In this project, we not only extend the studies on disruption management with fixed job processing times to the case with variable job processing times, but also propose integrated optimization methods of resource allocation, preventive maintenance, and job rejection with rescheduling. The research of this project will theoretically solve the disruption management problem for production scheduling with variable job processing times, enrich multi-objective evolutionary algorithm, and provide theoretical and methodological support for production enterprise to handle disruptions.

英文关键词： disruption management;integrated optimization;multi-objective evolutionary algorithm;surrogate model;apriori knowledge