论文部分内容阅读
神经树网络模型已成功应用于解决各类复杂的非线性问题,并且神经树网络模型的优化过程一般是先拓扑结构优化再参数优化,这种无参数信息的结构优化策略的缺点是干扰适应度的评价.鉴于此,提出一种改进的遗传规划(BGP)算法来综合神经树网络模型的两个优化过程.在两个时间序列预测问题上的仿真实验结果表明,所提出算法是一种有潜力且具备较好效能的方法.
The neural network model has been successfully applied to solve all kinds of complex nonlinear problems, and the optimization process of the neural network model is generally the first topology optimization and then the parameter optimization. The disadvantage of the structure optimization strategy without parameters is the interference fitness In view of this, an improved genetic algorithm (BGP) algorithm is proposed to synthesize two optimization processes of neural network model.The simulation results on two time series prediction problems show that the proposed algorithm is a Potential and have better performance.