多目标遗传算法求解模糊柔性作业车间调度问题

Multi-objective Genetic Algorithm to Solve Fuzzy Flexible Job Shop Scheduling Problem

建立了模糊数表示工件加工时间和交货期,以最大化平均客户满意度和最小化模糊完工时间为优化目标的多目标模糊柔性作业车间调度问题(Multi-objective fuzzy flexible job shop scheduling problem,MOFFJSP)模型,提出一种改进的多目标遗传算法(Multi-objective genetic algorithm,MOGA)。使用基于工序的编码方式表示调度解,并采用活动化解码方法改进解的质量。采用基于免疫和熵原理设计的交叉和变异算子来构造新解,并利用非支配排序和改进的精英保留策略来提升帕累托解集的多样性。通过仿真实验证明,改进的MOGA能够有效求解MOFFJSP问题模型。

A multi-objective fuzzy flexible job shop scheduling problem(MOFFJSP) is established that uses fuzzy numbers to represent workpiece processing time and delivery date,and the optimization goal is to maximize average customer satisfaction and minimize fuzzy completion time.Process-based encoding is used to represent scheduling solutions,and active decoding methods to improve the quality of the solutions.The algorithm uses crossover and mutation operators designed based on the principles of immunity and entropy to construct new solutions,and uses non-dominated sorting and improved elite retention Strategies to improve the diversity of Pareto solution sets.Simulation experiments prove that the proposed MOGA can effectively solve the problem model.