Solution by Genetic Algorithm In Regard to Sequence-Dependent Setup Time for Unrelated Parallel Machines Scheduling Problems

• Main Authors: Jeng-Wei Tseng, 曾政偉
• Other Authors: Pei-Hsi Liu
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/61545397683774752374

碩士 === 國立勤益科技大學 === 工業工程與管理系 === 95 === Some of the factory machines in these research cases are to be upgraded with new equipments, therefore, in certain work assignment, the machine configuration changes, i.e. new and old machines coexist together in particular production process. Some of the work machines are definitely performing faster than others. And certain manufacturing process is restricted to be carried out on certain machines. Accordingly, this research focuses on unrelated parallel machine scheduling problems and takes into considerations to minimize the overall delay for processing orders and the total production time. Another individual case needed to be considered is that, since the production materials are varied in nature, and the molds replaced must be cleaned prior production line change. As a consequence, the preparation time for every production line must be included as such. Subsequently, a parallel machine scheduling system configured on genetic algorithms can be used to find solutions according the superior search capabilities of the former. The parameter settings in genetic algorithm are critical to solution’s efficiency and effectiveness. Hence, Taguchi method is used to determine the parameter design and fine-tuning in the process of executing genetic algorithm, and eventually the most optimum parameter combination can be located. During this study, a case of manufacturer for lamps and lanterns was adopted to prove the effectiveness and stability of Taguchi method. And further comparisons were made against other frequently used traditional work scheduling methods. The final findings prove that, the configuration system proposed by this research can acquire fairly good scheduling results under the production environment from companies adopted as study case.