The Application Of Genetic Algorithm To The Match Of No-Entrusted Slabs And Contracts

The match of non-entrusted slab and contract optimization problem comes from modern steel industry, which is a non-linear and multi-objective large-scale optimization problem with constraint. Due to its complexity lies in multiple mutual conflicting problems and huge search spaces with high dimensions, it is difficult to work out satisfactory solutions using traditionally precise algorithms in real time and require to constructing highly effective optimization strategy.So this paper investigates the multi-objective optimization principle and several important traditional algorithms, discusses concerned important searching strategies about the multi-objective genetic algorithm and constructs math model of optimization match through deeply analyzing actual problems.On this basis, the paper adopts traditional multi-objective weighting coefficient method comparatively popular in industry fields and NSGA-II algorithm respectively, in which the adaptive genetic strategy based on individual-restoring means are used to meet the constraint condition and ensure the convergent speed. Comparatively speaking, both algorithms have their own advantages. Weighting coefficient algorithm is simple, speedy but can only obtain a solution once a time and requires decision-makers to set difficult weighting factors with enough experiences in order to achieve difficult solutions under difficult needs. NSGA-II algorithm need not parameter setting, could directly optimize and gain a group of Pareto solutions once a time which can be chosen according to difficult favors of decision-makers, but is complicate and time-exhausting.In order to test the quality of the optimizing algorithm and make an objective and fair judgment, the paper tentatively studies out the match optimization evaluating index. Both algorithms are applied in the match model of no-entrusted slabs and contracts in Shanghai Baosteel enterprise. Through analyzing and judging the match results, the paper proves the algorithms achieves comparatively ideal emulation result, realizes optimization match and thus validates to be adaptive. The research establishes theory basis for the algorithms being applied in the real manufacturing and provides a new approach to match the slabs and contracts more scientifically and more reasonably.