基于自适应遗传算法和BP神经网络的云服务器请求量预测模型

Cloud Server Request Volume Prediction Model based on Adaptive Genetic Algorithm and BP Neural Network

针对现有弹性云服务器(Elastic Cloud Server,ECS)未来请求量预测模型准确度低,稳定性差等问题,提出了一种基于自适应遗传算法和BP神经网络的预测模型。该模型以BP神经网络作为基础模型进行预测。采用自适应遗传算法(Genetic Algorithm,GA)优化神经网络初始权值和阈值,防止BP神经网络训练过程中陷入局部极小值。在自适应遗传算法初期引入多子代交叉方法加快遗传算法的收敛速度。通过对比实验表明,该模型在实际云服务器请求量预测过程中具有更好的准确性和稳定性。

Aiming at low accuracy and poor stability of the future request volume prediction model of ECS (Elastic Cloud Server), a prediction model based on adaptive genetic algorithm and BP neural network is proposed. The model, with BP (Back Propagation) neural network as the basic model for prediction, adopts the adaptive genetic algorithm (GA, Genetic algorithm) to optimize the initial weight and threshold of the neural network, thus to prevent the BP neural network from falling into local minimum values during the training process. The multi-child cross method is introduced in the early stage of adaptive genetic algorithm to accelerate the convergence speed of genetic algorithm. The comparison experiments indicate that this model has better accuracy and stability in the actual cloud server request volume prediction process.