摘要
提出了求解不等式约束非线性优化问题的群体复合形进化算法,提出的算法能充分利用目标函数值的信息、优化搜索过程具有较强的方向性和目标性,收敛速度快,且是全局优化算法;将群体复合形进化算法应用于三层前向人工神经网络逼近,提出了三层前向人工神经网络全局最优逼近算法;将三层前向人工神经网络全局最优逼近算法应用于实例,表明了提出的全局最优逼近算法的有效性。
In this paper,a more effective method that called multi-Complexes Evolution Algorithm for Constraint Nonlinear Optimization Problems ispresented.This new method can use the information of the objective function;the searching process for optimization of the method is directional and objective.This new method is a global optimization method.Multi-Complexes evolution Algorthm is applied to three-layer feed forward artificial neural network,and the algorithm of three-layer feed forward artificial network global optimum approach is presented.Several cases are studied;the validity of the algorithm of three-layer feed forward artificial neural network global optimum approach is confirmed.
关键词
前向神经网络
全局优化算法
进化算法
复合形算法
feed forward artifictial neural Network
Global Optimization Method
Evolution algorithm
Complexes Algorithm.
