Research On Scheduling Semiconductor Manufacturing Using Genetic Algorithm

Semiconductor manufacturing is one of the most complicated manufacturing systems due to re-entrant process flows, complex technology and equipment, high uncertainty and so on. Its scheduling is very important to improve the productivity of the whole manufacturing line. The semiconductor manufacturing scheduling problem is NP-hard, and has received considerable attention in academia and industry. This paper considers two scheduling problems in the semiconductor manufacturing industry and designs two genetic algorithms for solving them, respectively.At first, this paper addresses the large-size multi-stage multi-product scheduling problem of parallel processing machines with changeover time between adjacent tasks processed on the same machine, aiming at minimizing the total flow time or total process time. We propose a genetic algorithm to solve this problem. Chromosome contains two gene segments, one gene segments is a task sequence and the other is a feasible machine sequence. Active scheduling technique and repair strategy are applied for increasing the search speed and finding good solutions. Computational experiment proves that the proposed genetic algorithm can get better solutionsThen, we consider the two-stage hybrid flow shop scheduling problem in which the first stage contains several non-identical discrete machines, and the second stage contains several non-identical batching machines. Each discrete machine can process no more than one task at time, and each batching machine can process several tasks simultaneously in a batch, as long as the total size of all jobs in the batch does not exceed the machine capacity. The processing time of a batch is represented by the longest time among all the jobs in the batch, and all tasks of the same batch start and finish together. The goal is to make batching and sequencing decisions in order to minimize the total flow time. Here the dynamic programming algorithm will be combined with genetic algorithm to search the optimal solution for this problem. Computational experiments indicate that the algorithm can have good performance.