关键词: 多小波, 阈值, 浮点数编码, 消噪变异, 遗传算法

Abstract: Coding issue is a difficult problem of Genetic Algorithm（GA） research. Floating Point Representation（FPR） is superior to other codes in function optimization and restriction optimization. FPR can improve local search performance of GA. Lately, the attention of researches is attracted to the noises generated by FPR in genetic environment and its influence on algorithm performance. But few research result is emerged on FPR denoising mutation based on multiwavelet thresholding. In this paper, the properties of multiwavelet and FPR noises are researched, the method of FPR denoising mutation is presented on multiwavelet thresholding, the experiments are carried out by the author. The experiment result is compared with other method. The results of the research and the experiments indicate which the method can raise the convergence precision and velocity of the algorithm, improve whole performance of the algorithm.

Key words: multiwavelet, threshold, Floating Point Representation（FPR）, denoising mutation, genetic algorithm