摘要
研究无线网络优化控制问题,针对流量拥塞,传统BP小波网络(BPWNN)易陷入局部极小,收敛速度慢的缺陷,为了提高网络服务性能,提出一种改时的学习速率自适应的算法IBPWNN。在IBPWNN的小波网络运行过程中,动态的调整学习速率,防止网络陷入局部极小和误差变大。进行仿真实验,分别利用BPNN和IBPWNN进行网络流量进行对比预测。仿真结果证明,IBPWNN算法既简捷,又能够提高学习速度和精度,避免了BPWNN网络易出现的收敛速度慢、易产生局部最优解的问题,为网络优化提供了有效算法。
A prediction model based on improved back propagation wavelet neural network is established for traffic congestion of wireless network optimal control problem. Considering that traditional back propagation wavelet neural network is easy to take local convergence and has slowly learning convergent velocity, a method based on adaptive learning rate is used to optimize it in accelerating the learning convergence velocity. To prevent the network into local minimum and the larger error, an improved back propagation wavelet neural network （IBPWNN） adjustments learn- ing rate dynamic in the wavelet network during training. The speed and prediction precision of IBPWNN is compared with that of back propagation neural network（BPNN） of in the case of network traffic prediction. Theoretical analysis and simulation result show that IBPWNN is an effective algorithm. It not only is simple and efficient like BPWNN algorithm but also can increase the learning speed and prediction precision, which avoids the problem that BP neural network is easy to be trapped in local minima and its convergence speed is slow. It is an effective network optimization algorithm.
出处
《计算机仿真》
CSCD
北大核心
2010年第6期187-190,共4页
Computer Simulation
关键词
网络流量
小波分析
神经网络
预测
Network traffic
Wavelet analysis
Neural network
Prediction
