Research On Bilateral-aware Proactive Service Recommendation In Mobile Edge Computing

With the rapid development and maturity of Mobile Edge Computing(MEC)technology,a large number of services from different fields have appeared on the network.At the same time,more and more users are increasingly rely on services to complete work and life.However,the proliferation of service type and quantities and the dynamic nature of Quality of Service(QoS)in MEC make it difficult for users to quickly find reliable services that meet their needs,which seriously affecting user satisfaction and service resources utilization.Therefore,the rapid discovery of reliable services in MEC has become one of the important challenges to be urgently addressed in the service field.Aming at this problem,this paper proposes a bilateral-aware active service recommendation method.This mothod firstly actively senses the user's service needs based on the improved online deep learning method.After that,it predicts the multi-dimensional QoS value of candidate services based on the multi-tasking deep learning method of multi-dimensional context awareness.Finally,according to the predicted QoS value,the active Services with excellent QoS evaluation values are recommended to users.This research work is intended to deepen the combination of artificial intelligence and service computing,and explore new theories and methods of intelligent services,which have important theoretical significance and use value.The main research contents of this paper are as follows.1.For the user side,an active user demand perception method based on Improved Online Deep Learning(IODL)is proposed.This method first designs a data driven Rectified Linear Unit(DRe LU).After that,it incorporates DRe LU into a hedge back-propagation online deep learning model.Finally,an online demand forecasting method based on the IODL model is proposed.The effectiveness of the proposed method is verified by a large number of experiments.2.For the service side,a multi-dimensional context-aware multi-QoS prediction method is proposed.Firstly,this method makes an in-depth analysis and modeling of the multi-dimensional context which have an important impact on the quality of service in MEC.After that,the concepts of data fusion and adaptive updating of weights of multi-task loss function are introduced into multi-task deep neural network(MTDNN),an improved multi-task deep neural network model(IMTDNN)is constructed and a multi-dimensional context-aware multi-dimensional QoS prediction method based on IMTDNN is given.Finally,a large number of experiments are carried out based on Edge Cloudsim platform to verify the effectiveness of the proposed method.