摘要
随着网络中大量功能相似的web服务不断涌现,服务质量(QoS)在web服务选择中的地位日益突出,但目前关于web服务QoS管理和动态预测的机制还不成熟.为了支持动态QoS预测,弥补当前web服务集成模型的不足,提出了一种面向QoS的多Agent web服务集成模型.然后在该模型的基础上,提出一种基于灰色神经网络的动态QoS预测方法,并运用MATLAB建立动态QoS预测模型,通过仿真验证本方法在QoS动态预测中是可行且有效的.
Along with emerge of a large number of web service in the network,quality of service（QoS）becomes more and more significant in the web service selection.The QoS prediction is not only an important means of web services selection,but also has an important significance of the entire service composition process.The Existing QoS prediction is divided into static QoS prediction and dynamic QoS prediction,and the existing QoS prediction method is divided into recommend prediction algorithm,reasoning prediction algorithm,artificial intelligence prediction algorithm.However,most of the prediction method is only for QoS static prediction problem of web service selection,lack of support QoS management and dynamic prediction model and prediction accuracy aspects to be improved.Software Agent is an independent function calculation entity of distributed system and the cooperation system,which has the characteristics of autonomy,interactivity,reactivity and initiative.In order to perform the dynamic QoS prediction and make up for the shortage of the current web service integration model,we put forward a web service integration model based on multi-agent called QoS oriented multi-agent web service integration model（QOMAWSIM）through joined the service agent and quality agent in the traditional web service layer.QOMAWSIM including application layer web service layer,intelligent Agent layer and the user layer.This model can provide QoS management operation,such as web service QoS validation,QoS negotiation and QoS monitoring,which makes it easier to realize the dynamic QoS prediction and dynamic service selection.Then based on QOMAWSIM,this paper put forward a new method for QoS prediction of web service using grey-neural network.Grey neural network not only have the advantages of accumulation generation in the grey forecasting method,but also play the neural network intelligent processing characteristic-cs.Finally MATLAB is used to build the model of QoS dynamic prediction and the simulation results show that this method is feasible and effective.
出处
《南京大学学报（自然科学版）》
CAS
CSCD
北大核心
2013年第2期268-277,共10页
Journal of Nanjing University（Natural Science）
基金
国防预研基金(AY208J003)
