| Summary: | 碩士 === 逢甲大學 === 統計學系統計與精算碩士班 === 106 === There have been many applications of two-stage three-machine assembly flow shop in query scheduling, such as fire engine assembly, personal computer manufacturing, and distributed database system. Moreover, learning phenomenon has been shown present in many two-stage assembly flow shop environments. In conjunction with this learning phenomenon, we addressed, in this study, a two-stage three-machine flow shop scheduling problem with a cumulated learning function. Our objective was to minimize the total completion time. Taking the average gap between a lower bound and near-optional solution as the criterion, we proposed six versions of hybrid particle swam optimization (PSO) algorithms and genetic algorithm (GA) for small-size and big-size jobs, and for three different data types. In addition, analysis of variance (ANOVA) was employed to examine the performances of the seven algorithms for each data type. Subsequently, Fisher’s least significant difference tests were carried out to further make pairwise comparisons among the performances of the seven algorithms.
Keywords: two-stage three-machine assembly; flow shop scheduling; hybrid particle swam optimization; genetic algorithm;total completion time; a cumulated learning function
|
|---|
