基于云自适应遗传算法的改进BP算法  

An Improved BP Algorithm Based on Cloud Self-Adaptive Genetic Algorithm

作　　者：吴立锋[1] 程林辉[1] 

机构地区：[1]中南民族大学计算机科学学院,武汉430074

出　　处：《中南民族大学学报（自然科学版）》2011年第4期98-101,共4页Journal of South-Central University for Nationalities：Natural Science Edition

基　　金：中南民族大学自然科学基金资助项目(YZQ09003)

摘　　要：针对标准BP算法对初始权值敏感、收敛速度慢、易陷入局部极小等问题,结合正态云模型云滴的随机性和稳定倾向性,以及遗传算法的全局搜索能力、收敛速度快等特性,提出了云自适应遗传改进BP算法.该算法首次将云模型和遗传算法结合调整神经网络的权值和阈值.由X条件云发生器产生改进的自适应交叉概率和变异概率.实验结果表明:云自适应遗传改进BP算法比标准BP算法收敛速度快.The standard BP algorithm is sensitive to the initial weights, converges slowly and is easy to trap into local minimum. Aiming at these limitations of the standard BP algorithm, combining the randomness and stability of the cloud droplets in the normal cloud model, and the global search ability and fast convergence of the genetic algorithm, the cloud self-adaptive genetic BP algorithm is put forward in this paper. This algorithm firstly combines the cloud model and the genetic algorithm to adjust the weights and threshold values of the neural network. The improved self-adaptive crossover probability and mutation probability are generated by X-conditional cloud generator. The results of the experiment show that the convergence speed of the cloud self-adaptive genetic BP algorithm is faster than that of the standard BP algorithm.

关 键 词：云模型 遗传算法 云自适应遗传BP算法 神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]