摘要
介绍了遗传算法和标准BP算法及其改进形式,指出遗传算法和BP算法各自的优缺点。利用遗传算法全局寻优和BP神经网络局部寻优相结合的方法,提高了传统BP神经网络的计算精度和收敛速度。最后进行了仿真实验,结果表明,该方法不仅收敛速度快,而且易达到最优解,具有很高的实用价值。
The genetic algorithm, standard back propagation and the corresponding improved methods are introduced with their merits and defects presented. By combining GA （Genetic Algorithm）, which has the advantage of global optimization＇, and BP（ Back Propagation） which has the advantage of local optimization, the calculation accuracy and convergence rate of the traditional BP neural network are improved. The simulation re- sult shows this method has high convergent speed and easily-oriented global optimization and is therefore of great practical value.
出处
《电子科技》
2008年第11期59-62,共4页
Electronic Science and Technology
关键词
神经网络
遗传算法
网络训练
反向传播算法
neural network
genetic algorithm
network training
back propagation algorithm
