将求解不等式约束非线性优化问题的群体复合形进化算法 ,应用于前向人工神经网络逼近 ,提出了前向人工神经网络全局最优逼近算法 ;将前向人工神经网络全局最优逼近算法应用于太湖水位预报 ,建立了太湖水位预报的神经网络模型 ,表明了提出的全局最优逼近算法的有效性 The complex compound evolutionary algorithm for solving nonlinear optimization problems with inequality constraints is applied to the approximation of forward artificial neural networks. A global optimal approximation algorithm of forward artificial neural networks is proposed. The global optimal approximation algorithm of forward artificial neural networks The Taihu Lake water level forecasting, the establishment of the Taihu Lake water level forecasting neural network model, shows the effectiveness of the proposed global optimal approximation algorithm