基于蝙蝠算法优化反向传播神经网络模型的无线网络流量预测

Wireless network traffic prediction based on bat algorithm optimized back propagation neural network model

针对无线网络流量数据预测精度不高问题,提出一种基于蝙蝠算法(BA)优化的反向传播(BP)神经网络的分类预测模型——BABP。通过采用蝙蝠算法对BP神经网络模型的初始权值与阈值进行全局寻优,构建崭新的基于蝙蝠算法优化的神经网络模型。通过与基于传统寻优算法遗传算法(GA)与粒子群优化(PSO)算法的反向传播(BP)神经网络模型比较,在无线网络流量数据的分类预测和稳定性方面,提出的BABP模型要优于GABP模型、PSOBP模型;同时,无论迭代次数的多与少,BABP均比GABP、PSOBP算法更快地收敛。实验结果表明,BABP模型在预测精度、寻优速度以及模型稳定性等方面均比GABP、PSOBP模型更具优势。

Aiming at the low prediction accuracy of wireless network traffic data,a classification prediction model named BABP(Bat Algorithm optimized Back Propagation),which is the Back Propagation(BP)neural network optimized by Bat Algorithm(BA),was proposed.By using bat algorithm to optimize the initial weights and thresholds of BP neural network model,a new neural network model based on bat algorithm optimization was constructed.Compared with the neural network models optimized by traditional Genetic Algorithm(GA)and Particle Swarm Optimization(PSO),the proposed BABP model is better than Genetic Algorithm BP(GABP)model and Particle Swarm Optimization BP(PSOBP)model in classification prediction and stability of wireless network traffic data;at the same time,regardless of the number of iterations,BABP has faster convergence speed than GABP and PSOBP algorithms.The experimental results show that BABP model has more advantages than GABP and PSOBP models in prediction accuracy,optimization speed and model stability.