提出了一种改进的自适应遗传算法 Ｉ Ａ Ｇ Ａ，它利用网络结构的特点，采用前向自适应技术，实现对神经网络的有效训练．实验表明，该算法优于 Ｂ Ｐ算法、标准遗传算法 Ｂ Ｇ Ａ 和普通自适应遗传算法 Ａ Ｇ Ａ，网络训练质量和效率都有很大提高 An improved adaptive genetic algorithm, I A G A, is proposed. It takes advantage of the characteristics of network structure and adopts forward adaptive technology to realize the effective training of neural network. Experiments show that the algorithm is superior to the B P algorithm, the standard genetic algorithm B G A and the general adaptive genetic algorithm A G A, the network training quality and efficiency are greatly improved