关键词: 工艺设计, 工艺路线优化, 遗传算法, 工步有向图, 拓扑排序, 约束矩阵

Abstract: Aming at process planning optimization in process design, the basic priority of constraints between operations were summed up, the entire process was transformed into an operation digraph based on the constraint relationship. The topological relationships among operation nodes were transformed into a constraint matrix, a detection approach for matrix-bound judgement was proposed. The operation graphs topological sort model was established. A random depth-first search algorithm to topologically sort out an operation digraph was designed and a certain number of feasible initial process planning qua the initial population of genetic algorithm was obtained. In the algorithm, processing sequence coding strategy based on the dynamic Job Shop resources was proposed. The evaluation criteria of the process planning was analyzed quantitatively, the objective function and constraints was established as a non-constraint optimization objective function by using penalty function, thus the chromosomes fitness function was determined. Genetic operations (selection, crossover and mutation) were developed. Algorithm astringency was speed up by elite tactic based on simulated annealing algorithm. Finally, an illustrative example was given to testify the effectiveness of this algorithm.

Key words: process design, process planning optimization, genetic algorithm, operation digraph, topological sort, constraint matrix