Algorithm Designing Of Varible Batch Flexible Job-shop Scheduling With Multi-products

With the developing of productivity, it is a high call that an enterprise improves its efficiency of producing. As batch processing of large scale is common today, researches focused on this field provide both theoretical and practical benefits.In this thesis, we have introduced the knowledge and method in scheduling problem in general, the research done by others in batch scheduling problem, the basic theory and procedure of genetic algorithm and its application in scheduling problems. We then treat the variable batch job shop scheduling problem using the intelligent searching algorithms (genetic algorithm, simulated annealing algorithm).Considering the realistic in manufacturing, we propose that divide the processing time into two parts: the machine-preparing time that can be done before jobs'coming and the real processing time, and prove the idea to be efficient. Based on the considering of cutting down machine-preparing time when some jobs of the same kind are processed consecutively, we design a genetic algorithm to solve our problem (MPGA). Further, we add refresh operation into the algorithm to avoid the disadvantage when using genetic algorithm.When the batch number is large enough, we use the strategy of batch splitting to make them divided into smaller batches called sub-batches. The strategy is proved to be efficient by many researchers and our work in this thesis is to find the best splitting result. We propose a new kind of simulated annealing algorithm called Boltzmann simulated annealing (BSA) to search for the optimal splitting. BSA and MPGA infect each other and will converge at the same time, providing the optimal splitting as well as the best scheduling result.Simulation result shows that this thesis could solve the variable batch scheduling problem of multi-product. Both MPGA and BSA are proved to be efficient.