基于遗传神经网络的网格资源预测模型.pdfVIP

计算机仿真 2012 年4 月 第29 卷第4 期 文章编号:1∞6 -9348(2012)04-ω43 -04 基于遗传神经网络的网格资源预测模型 白杨 (呼伦贝尔学院，内蒙古呼伦贝尔 021∞的 摘要:研究网格资源预测问题，网格资源具有非线性、混沌变化特点，传统 BP 神经网络具有局部极小、收敛速度慢等缺陷， 预测精度较低。为提高了网格资源预测精度，提出一种基于遗传神经网络的网格资源预测模理。利用遗传算法对 BP 神经 网络的权值和阔值进行优化，然后采用 BP 神经网络对网格资源建立预测模型，最后采用网格资源时间序列进行有效性仿 真。仿真结果表明，遗传神经网络有效地解决了传统BP 神经网络的不足，提高了网格资源的预测精度，降低 f预测误差， 十分适合于非线性、混沌的网格资源时间序列预测。 关键词:神经网络:遗传算法:网格资源;预测 中图分类号:1凹91 文献标识码:B Prediction Model for Grid Resources Optimized BP Neural Network Based 00 Genetic Algorithm YANG Bai ( Hulunheier college , Hulunheier Neimenggu 但1∞8 ， China) ABSTRACT: Study grid re回urce prediction problem. Grid source is of the characteristics of nonlinear and cha俑， and the traditional BP neural network wi由 local minima , slow convergence speed and other defects has low forecas- ting precision. In order to improve the prediction accuracy of grid resources , this paper propo甜d a grid re酬lTCe pre- diction model based on the genetic neural network. The weights and thresholds of BP neural network WeI曹 optimized by genetic algorithm , Then the BP neural network wωu田dωestablish the prediction model of 回d source. Finally , the simulation experiments were carried out with grid source time seri回.也e simulation results show that the 阴阳ed method effectively improves the deficiency of traditional BP neural network and the prediction accuracy of grid re- ωurce ， and reduces the prediction error. It is very suitable for grid resources prediction 叨th nonlinear , cha帽 char­ actensbc. KEYWORDS: Neural network; Genetic algorithm; Grid resource