Research On Multi-objective Flexible Job-shop Scheduling Problem Based On Genetic Algorithm

Manufacturing industry of China is facing a huge challenge in the trend of global development. It is inevitable to improve efficiency of production and resource utilization to promote the transformation and upgrading of this industry. As flexible job-shop scheduling problem is a classic job shop scheduling problem, FJSP with optimizing every objective considering practical production scheduling is a much harder NP problem.In previous researches, the process time is considered constant, which is volatile in the actual production situation. Hence, based on literature review, this dissertation is going to improve genetic algorithm.While process time is considered to be constant value in classic model of scheduling model, in this dissertation, flexible job-shop scheduling problem under multiple-objective optimization is researched where process time is exponential distribution. The targets is to balance workload of machines and reduce delays. An effective genetic algorithm is proposed in this dissertation which initializes by OOC(objective oriented coding) and RC(random coding) in order to facilitate the algorithm efficiency and keep population diversity. In order to searching the better optimal solution faster, fitness linear transformation and self-adaptive crossover and mutation is employed to keep population diversity and to improve searching efficiency.After verification, this dissertation proposes an improved genetic algorithm. The effectiveness and feasibility of the algorithm in FJSP has been shown in an instance with multiple-objective optimization. While it is shown the search effectiveness and constringency speed is superior to self-adaptive genetic algorithm and standard genetic algorithm, which provide guidance for the future manufacturing production activities.