人工神经网络解决问题的方式不同于传统逻辑思维的“算法”,其“操作”具有形象思维的属性,特别适合解决数学模型难以描述的非结构化问题和非线性问题,具有良好的映射逼近能力,良好的可靠性、鲁棒性和容错性。误差反向传播的前馈式人工神经网络[1]是目前使用最广泛的一种人工神 The way to solve the problem by artificial neural network is different from the “algorithm” of traditional logical thinking. Its “operation” has the property of image thinking. It is especially suitable for solving the unstructured and nonlinear problems difficult to describe by mathematical models and has good mapping approximation ability , Good reliability, robustness and fault tolerance. Feed back error artificial propagation neural network [1] is currently the most widely used artificial God