自适应遗传算法优化管网状态估计神经网络模型

Self-adaptive GA for optimization of neural network model of pipe network for state estimation

针对遗传算法收敛速度慢、传统BP算法易收敛于局部最优以及网络结构难以确定等问题,引进自适应遗传算法优化网络的权阈值以搜寻网络最优拓扑结构,并利用自适应遗传算法优化该网络的权阈值,建立基于改进BP网络的宏观管网状态模型.实例分析表明,改进模型具有较高的预测精度.

Simple GA has the shortcoming of slow convergence, while BP neural network is prone to the local optimum and its structure is usually difficult to be determined. In this paper, a real-coded self-adaptive GA was introduced to optimize the weight and threshold so that a binary-coded self-adaptive GA can find the best topological structure, and then a real-coded self-adaptive GA was presented to optimize the weight and threshold of BP network which has the best structure. A macroscopic state model of pipe network was developed based on improved BP neural network. Case study shows that the improved model has higher prediction accuracy.