求解模糊作业车间调度问题的群体邻域搜索算法

摘要

This paper presents a swarm-based neighbourhood search algorithm(SNS) to minimize the maximum completion time of fuzzy job shop scheduling problem(FJSSP). SNS uses an ordered operation-based representation, tournament selection and swap operation in which probability is 1 and swap-time is adjusted dynamically. The theoretical analyses on scheduling results show that fuzzy makespan can be used to evaluate the quality of solution. Theoretical analysis and a large number of experiments demon strate that SNS has strong global and local optimization capabilities,faster convergence speed,and promising performance on FJSSP.%本文提出一种群体邻域搜索算法(Swarm-based Neighborhood Search,SNS),用于最小化模糊作业车间调度问题(Fuzzy Job Shop Scheduling Problem,FJSSP)的模糊makespan.该算法使用基于有序工序的编码,通过锦标赛选择和概率为1的动态调整互换操作更新群体.对调度结果的理论分析表明,模糊makespan能反映解的优劣.理论分析及大量实验证明,SNS具有较强的全局和局部优化能力,以及较快的收敛速度,在求解FJSSP方面具有较强的优势.