对电磁兼容进行预测采用传统的BP神经网络易于陷入局部最优,为了解决上述缺陷,本文采用遗传算法对网络权值进行优化。以平行线间电磁耦合干扰为具体算例,证明本文算法的预测结果的均方误差仅有10-4数量级。本文提出的用遗传算法优化网络权值的方法有效,且神经网络模型能准确预测电磁兼容。 The electromagnetic compatibility prediction using traditional BP neural network tends to fall into the local optimum, in order to solve the above shortcomings, this paper uses genetic algorithms to optimize the network weights. Taking the electromagnetic coupling interference between parallel lines as a concrete example, it is proved that the mean square error of the prediction results of this algorithm is only 10-4 orders of magnitude. In this paper, the genetic algorithm to optimize the network weights is effective, and the neural network model can accurately predict the electromagnetic compatibility.