GEP算法在无线传感器网络数据预测中的仿真应用

Simulation Application of GEP in WSNs Data Forecasting

数据预测是研究无线传感器网络数据融合的重要方法。利用因特尔伯克利实验室开源数据为样本,采用时序预测方法分析了数据的时间相关性,深入研究了基因表达式编程方法(GEP)在无线传感器网络预测应用的可能性。通过与同等条件下ARMA算法进行比较分析可知,GEP算法不依赖先验知识,并且预测精度高于ARMA算法,为GEP算法在无线传感器网络数据融合应用提供了研究依据。

Data forecasting is an important method of data fusion to study wireless sensor networks. Intel-Berkeley Lab open resource is chosen as data sample,and temporal correlation is analyzed by using time series forecasting. Especially,the proposed approach relies on autoregressive models built at each sensor to predict local readings,which is built by GEP algorithm. GEP algorithm does not require priori knowledge in comparison with ARM A algorithm under same condition,which has higher precision.