基于GSO优化BP神经网络的预测模型

Predictive models of optimizing BP neural network based on GSO

针对BP神经网络及人工萤火虫群智能算法(Glowworm Swarm Optimization,简称GSO)两种智能算法的不足,提出了一种利用GSO优化BP神经网络的新算法。该算法主要是利用人工萤火虫算法训练BP神经网络,以此来提高混合算法的精确度。仿真试验结果表明,利用GSO算法优化BP神经网络可以有效地提高预测变形的精确度。

A new algorithm optimizing BP neural network based on GSO is proposed aimed at the lack of algo-rithm for both BP neural network and Glowworm Swarm Optimization （GSO）. The algorithm is to use GSO to train BP neural networks in order to improve accuracy of hybrid algorithm. Simulation results show that using GSO algorithm for optimizing BP neural network can improve the accuracy of predicted deformation.