具有多处理机任务约束的混合作业车间调度建模与嵌入式仿真

项目名称： 具有多处理机任务约束的混合作业车间调度建模与嵌入式仿真

项目编号： No.71502015

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： 樊坤

作者单位： 北京林业大学

项目金额： 18万元

中文摘要： 随着多品种、小批量逐渐成为离散制造企业的主要生产模式，如何对具有多处理机任务的作业车间进行高效排产已成为许多离散制造企业亟待解决的问题，同时也逐渐成为国内外学者研究的热点。本项目针对具有多处理机任务约束的混合作业车间调度（HJSMT）问题进行研究。首先针对确定性HJSMT问题，设计出基于粒子群算法的启发式求解算法。由于在实际生产中会具有许多不确定性，使得工件加工时间无法精确预知，因此本项目假设工件具有随机加工时间，采用嵌入式仿真技术对随机HJSMT问题进行仿真求解时，利用基于粒子群的启发式算法进行任务分配，得到的调度仿真结果可为实际调度提供重要和有意义的参考。抗风险是调度决策者最常考虑的因素，本项目采用Scenario方法来描述随机加工时间，建立期望-方差模型和鲁棒优化模型并进行求解，使得到的调度决策不仅具有较好的期望值，而且还具有一定的抗风险能力，为风险规避型决策者提供最优调度风险决策。

中文关键词： 生产计划与调度；作业车间；多处理机任务；粒子群优化；嵌入式仿真

英文摘要： With the multi-species and small batch becoming the main production mode of discrete manufacturing companies, job shop with multiprocessor tasks for efficient scheduling is an urgent problem of these enterprises to be solved and is gradually becoming the hot study issue of researchers. The project will study the Hybird Job-shop Scheduling with Multiprocessor Task (HJSMT) problem. .Firstly, a heuristic algorithm based on Particle Swarm Optimization (PSO) is developed to solve the deterministic HJSMT problem. Since there are many uncertainties in the actual production, the processing time of jobs cannot be accurately predicted. Assuming the jobs with random processing time, the embedded simulation technology is used to simulate the stochastic HJSMT problem and the heuristic algorithm based on PSO is proposed for task allocation. The simulation results can provide an important and meaningful reference for the actual production scheduling..The anti - risk is most often considered by scheduling policy makers, so we use the Scenario method to describe the random processing time and establish expectations - variance model and robust optimization model of stochastic HJSMT. The solution of the robust optimization model not only has good expectations, but also has a certain ability to resist risks, and it can provide the optimal scheduling risk decisions to the risk avoidance decision-makers.

英文关键词： Production planning and scheduling;Job-shop;Multiprocessor task;Particle swarm optimization;Embedded simulation