Research on Threshold Denoising of FPRGA

Abstract

Genetic algorithm (GA) was widely used to engineering and optimization fields. The performance of GA and extension of its application fields were affected by the limitation of its code. Floating point representation (FPR) is super to other codes in function and constraint optimization fields. Noise was brought by the FPR in selection and crossover operation,its influence to the performance of GA was not noticed by re searchers. In this paper, the property of the noise in FPR is mostly analyzed in inherit operation. The mechanism of denoising in FPR is researched with wavelet threshold coefficient. Denoising is implemented by mutation. That wavelet theory used in floating point representation genetic algorithm (FPRGA) is credible for reducing noise level, the method is feasible, is indicated by its results of research and experiment.