基于模糊进程因子的改进自适应遗传算法

An Adaptive Genetic Algorithm Based on the Fuzzy Process Factor

首先提出了一种改进自适应遗传算法,在本算法中引入模糊进程因子,设置了最佳样本保留数、最佳基因保留数等五个可调参数,使它们随着模糊进程因子的变化而变化,从而实现算法自适应;然后给出了新算法的步骤,并用标准测试函数对该算法进行了测试,测试结果表明:该算法具有较好的在线性能,提高了收敛速度;最后,用该新算法进行了功率谱估计,该算法比LS法具有更好的谱分辨率和适应性。

An improved adaptive genetic algorithm is proposed in this paper. In this new algorithm, a fuzzy process factor is introduced to the operations of crossover, mutation and selection in the standard genetic algorithm. Five adjus table parameters such as the best individuals keeping number are set in the algo rithm. These parameters are varied with the fuzzy process factor, so that the se lf-adaptability of the algorithm is realized. The algorithm procedures are give n in detail. It is tested using the standard test functions. The results show th at the new algorithm has good online performance and fast convergent velocity. A t last, using the improved genetic algorithm to estimate the power spectrum unde r the criterion of minimizing error prediction. The power spectrum estimation me thod based on the improved genetic algorithm has better spectrum distinguishing rat and fitness than LS method.