A New Genetic Algorithm for the Order Acceptance Single and Parallel Machine Scheduling Problems

• Main Authors: Chang, Yu-Tang, 張鈺堂
• Other Authors: Chen, Shih-Hsin
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/n9ay25

碩士 === 正修科技大學 === 資訊管理研究所 === 105 === Scheduling problems exist in many industries, including the processing schedule of the printing industry, the flight attendant flight schedule, and the hot rolling of the steel. In this study, we focus on the order acceptance scheduling problems (OAS) on a single machine and parallel machines, which are an extension of the scheduling problem. OAS problems will be much more practical because a firm may not accept all the orders due to the limitation of production capacity which maximizes the total profits. Some orders are accepted while others might be rejected because these orders do not meet the deadline. In our studied problems, we consider a setup time for each single and parallel machine scheduling problem. In particular, a past-sequence-dependant (PSD) is introduced which has not been studied yet. To solve the different scheduling problems effectively, a two-part encoding genetic algorithm is proposed together with a new crossover operator and a local search operator from literature. The results show the local search operator is useful regarding the solution quality and the CPU time. Hence, it might point out a research direction for the researchers or practitioners.