关键词: 冷轧机组, 多背包问题, 系统聚类法, 自适应量子遗传算法

Abstract: To solve the problem of coils merging in the cold rolling mill, a multi-objective and multi-knapsack model was established by considering the merging mode and the maximal available coils weight. The objective of this model was to maximize the number of merging and to minimize the deviation of the coil's weight simultaneously. Based on this model, firstly the system clustering method was used to acquire the classification of coils and the knapsack model's center so as to simplify the model's computation. Then a self-adaptive quantum genetic algorithm was developed to solve the coils merging model, in which the quantum door revolving gate angle and no-door mutation probability were adaptively adjusted according to the state of population's concentration-dispersion and evolution. The designed self-adaptive adjustment mechanism ensured the algorithm's search ability and speeded up the convergence at the same time. Finally, the simulation results using the classical knapsack problem and practical production data showed that the proposed model and algorithm were feasible and effective.

Key words: cold rolling mill, multi-knapsack problem, system clusters method, adaptive quantum genetic algorithm