HYBRID GENETIC ALGORITHMS FOR MINIMIZING THE MAXIMUM COMPLETION TIME FOR THE WAFER PROBING SCHEDULING PROBLEM

作     者：Yang, Ming-Hsien Lo, Chien-Chung Chen, Hsin-Min Huang, Hau-Lieng Ming-Hsien Yang;Chien-Chung Lo;Hsin-Min Chen;Hau-Lieng Huang

作者机构：Natl United Univ Dept Business Management 1 Lien Da Miaoli 36003 Taiwan 

出 版 物：《JOURNAL OF INDUSTRIAL AND PRODUCTION ENGINEERING》 (工业工程学刊)

年 卷 期：2005年第22卷第3期

页      面：218-225页

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 08[工学] 

基　　金：National Science Council  Taiwan  R.O.C. [NSC 92-2213-E-239-008] 

主　　题：hybrid genetic algorithms initial population minimum make span sequence dependent setup time due dates 

摘      要：The wafer probing scheduling problem (WPSP) is a variation of the parallel-machine problem, which carries with the objective to minimize the total workload to enhance the utilization rate of machine capacity. However, to minimize the maximum completion time for the wafer probing scheduling problem (WPSP) is very important for equally utilizing the capacity of parallel-machine, while satisfying the requirements of product types, product family, sequence dependent setup time, product-type dependent processing time, due dates of jobs, and capacities of machines. To solve such a complicated problem efficiently and effectively, we adopt the hybrid genetic algorithm approach, which generates initial population with the insertion and savings algorithms for WPSP and provide a new crossover operator to generate better solutions. Considering the disturbance of due dates of jobs in scheduling, we design a new crossover which makes each machine keep the critical sub-schedule in the generation of offspring. The performance of the proposed hybrid genetic algorithm is evaluated by two real-world problems and the solution quality of the proposed approach is analyzed under five levels of mutation rates and four levels of number of generations. Computational results reveal that the hybrid genetic algorithm can efficiently solve the considered problem and the reduction of makespan is related to the mutation rates and the number of generations.