BP neural network optimize based on improved genetic algorithm

It is known that the single genetic algorithm (SGA) has many disadvantages, and the paper presents an improved genetic algorithm, which with a new genetic algorithm based on the fitness values and group diversity to optimize the BP neural network. Experiment has shown that the improved genetic algorithm cannot only solve the problems of initializing the group fitness exception, but also can various the groups by calculating the similarity in algorithm, to avoid premature convergence of the algorithm, and then accelerate the speed of learning convergence, made the generalization ability of neural network improved has a certain prospect in practice.