基于Epsilon约束法的短纤维生产调度的双目标优化研究

针对短纤维生产行业实际,本文综合考虑客户的需求差异、客户的重要程度、纤维生产设备的准备时间以及交货期差异等因素,研究连续需求下的短纤维生产排序优化问题。首先,本文建立双目标整数规划模型,即最小化客户订单总延迟和最小化机器总准备时间;其次,设计Epsilon约束算法并调用CPLEX精确求解调度方案,即帕累托前沿;最后设计非支配排序的遗传算法(NSGA-II)求解大规模生产下的调度优化方案。通过实验,证明该整数规划模型和算法对解决多客户连续需求问题具有实际价值,进而可以为短纤维生产企业提供参考。

An Exact Epsilon-constraint Algorithm for the Bi-objective Optimization Problem of Scheduling Staple Fiber Production

This paper studies the scheduling problem of staple fiber production with successive orders, which considers the demand and due date differences between clients and the setup time of machine when switching components. Firstly, we establish a bi-objective integer linear programming model so as to minimize the total tardiness of orders and total setup time of machine. Then, we conduct an accurate algorithm to get the precise solutions of proposed model, i.e., Pareto frontier. Finally, we NSGA-II to sovle this model within a devise substantial amount of time. Simulation experiments prove the correctness and effectiveness of this model and proposed algorithms, which can be considered to use as an reference in daily process.