Application of Modified Genetic Algorithms on Job-Shop Scheduling Problems

• Main Authors: Yi-Yuan Li, 李宜原
• Other Authors: Hsin-Chuan Kuo
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/37445716884183335727

碩士 === 國立臺灣海洋大學 === 系統工程暨造船學系 === 92 === ABSTRACT In recent years, the heuristic algorithm has gradually been adopted to deal with NP-hard problems in the combinatorial optimization. Among these combinatorial optimal problems, the job-shop scheduling problems are usually encountered in the production managements. In this thesis, the modified genetic algorithms (MGA) are proposed to analyze and further apply to the job-shop scheduling problems for improving the calculation effectiveness and efficiency. First of all, discussions on the operation procedures, restrictions, dispatching and effectiveness evaluating criteria of the job scheduling problems are made. Then, the fundamental framework and some important factors are introduced, and the coding and the cross-over formats of the GA are identified. After introducing several improvement strategies into the searching solvers, an MGA is formulated and used to compare the searching efficiency with the conventional GA. A series of testing examples on the controlling factors, such as cross-over rate, mutation rate and population size, are also conducted to analyze the influences of three different crossover formats on the conventional GA. Moreover, the solving and searching efficiencies of the MGA applied to the job-shop scheduling system are also discussed and compared with the traditional GA. Keywords: Job-shop scheduling problems, genetic algorithm, modified genetic algorithm, crossover rate, mutation rate, population size.