求解具前置工作站之分散彈性零工式排程問題

Abstract

本研究求解具前置工作站的分散且彈性的零工式排程問題 (Distributed and Flexible Job Shop Scheduling Problem, DFJSP)。各工件必須先經過前置工作站後再進行後續加工，前置工作站之工件排序對於整體的排程系統表現有顯著的影響，因此，此排程問題包含四個排程決策，(1)工件指派 (Job-to Cell Assignment)，(2) 作業指派 (Operation-to-Machine Assignment)， (3) 作業排序 (Operation-Sequencing for Machines) 和 (4) 前置工作站之工件排序 (Job Sequence on the Preprocess Workstation)，此研究應用Wu et al. (2015)發展出新的染色體表達之基因演算法，又稱S_GA_OP，以過去學者提出之S_GA_JS和S_IGA做比較，結果顯示S_GA_OP績效較佳。

This research studies a scheduling problem for a distributed flexible job shop which involves a preprocess workstation. That is, each job has to go through the pre-process workstation and proceed to the distributed flexible jobs shop system. The job sequence of the preprocess workstation surely has a significant impact on the scheduling performance. Therefore, the scheduling problem involves four scheduling decisions: (1) job-to-cell assignment, (2) operation-to-machine assignment, (3) operation sequencing, and (4) job sequence on the preprocess workstation. This study develops a new chromosome representation which is adapted from Wu et al. (2015) and proceeds to develop a genetic algorithm called S_GA_OP. Two other algorithms (called S_GA_JOB and S_GA_IGA) which are adapted from prior studies are taken as benchmarks for comparison. Experiments reveal that S_GA_OP outperforms the other two algorithms.