I use genetic algorithm to optimize the initial weights of the neural network, genetic algorithm (ga) is the error of the neural network calculation value, the fitness of the iterative evolution, select the smallest error value individual corresponding initial weights and bias, and then to the neural network calculation, the results with genetic algorithm iteration of error deviation, such as my genetic algorithm in the optimization to the fitness of a few representative error, then the individual sequence into neural network computation, can not meet a little a few error, how be to return a responsibility excuse me,