基于遗传算法神经网络盲均衡算法的研究

将遗传算法引入神经网络盲均衡,利用其全局搜索能力强的特性来消除传统神经网络算法易陷入局部最优解、训练速度慢的缺点。采用两阶段寻优法,首先,通过遗传算法来为神经网络提供一个全局较优的局部搜索空间;其次,利用传统神经网络在这个局部空间进行更精确地搜索,最终实现盲均衡。计算机仿真表明,该算法能达到更好的收敛特性和均衡效果。

Blind Equalization Algorithm Based on Neural Network and Genetic Algorithm

In this paper,genetic algorithm(GA) that is good at global searching is used in the blind equalization based on neural network.It can avoid the situations that are liable to get into the local minimum and slow speed in training of conventional neural networks.It is a two-step method.First,a preferable local solution space is offered to the neural network by using GA.Second,a precise searching is realized in the space with conventional neural network algorithm.And the blind equalization is finished.The computer simulations show that the algorithm obtains a good convergence characteristic and satisfied equalization result.