云计算中基于IABC算法的负载预测的研究

The Research on the Load Prediction of Cloud Computing Based on IABC Algorithm

摘要:

针对云计算中的资源负载预测的问题,采用改进的人工蜂群算法和SVM相互结合的方式构建预测模型,首先通过采用反向学习对种群进行初始化,差分进化对种群个体进行选择,吸引点策略构建算法的蜜源选择,通过反馈机制和森林法则降低算法陷入局部最优的缺点,其次,将SVM预测模型中参数通过改进的蜂群算法进行优化得到了最佳的参数,最后仿真实验中,提出的IABC算法预测精度优于SVM,LSSVM等预测算法,具有一定的推广价值。

Abstract:

Aiming at the problems about resources load prediction appear in the cloud computing, the IABC algorithm is used with the combination of SVM to construct model of prediction. Firstly, backward learning is used for population initialization, the individuals of the population should be selected by differential evolution, constructing the nectar selection of the algorithm by the strategy of attraction points, and decreasing the disadvantage of being trapped into partial bests by feedback mechanism and the law of the jungle. Secondly, refining the parameters in SVM prediction model by IABC algorithm to have best parameters. Lastly, in simulation experiences, the prediction of IABC algorithm is better than that of predicting algorithm like SVM, LSSVM and so on. So, the prediction of IABC algorithm is worth being promoted.