基于ACO-BPNN的话务量预测模型

Traffic forecasting based on ACO-BPNN

为了提高话务量的预测精度,针对传统人工神经网络存在的参数优化问题,提出一种蚁群算法优化神经网络的话务量预测模型(ACO-BPNN)。首先考虑话务量数据之间的时间相关性,重构话务量的学习样本,然后将样本输入到神经网络进行训练,并通过蚁群算法不断优化神经网络连接权值和阈值,使实际输出与期望输出之间的误差最小,建立话务量预测模型,最后将预测模型用于某市移动通信网络的忙时话务量预测中。实验结果表明,本文模型提高了话务量的预测精度,尤其是多步预测效果明显优于对比模型,具有良好的实用性。

In order to improve the forecasting accuracy of traffic,according to the parameter optimization problem of traditional artificial neural network,this paper proposes a forecasting model of traffic based on ant colony optimization algorithm and neural network.Firstly,the temporal correlation between traffic data is considered and reconstruct learning sample of traffic,secondly,the samples are input to the neural network for training and ant colony optimization algorithm is used to optimize connection eights and threshold of neural network to keep error minimum for output and the expected output and build traffic forecasting model,finally traffic forecasting forecast model is test by a city mobile communication network date.The results show that the proposed model improves the forecasting accuracy of traffic,especially multi-step forecasting effect is better than the other models,which is valuable for improving call quality.