Research On Collaborative Cache Scheme Of Mobile Edge Computing

With the increasing number of user terminals and the development of 5G technology,a network has been arisen where macro base stations and small base stations co-exist.Meanwhile,applications such as ultra-high resolution video and VR/AR have higher requirements for latency.As an important technology in edge computing,edge caching can effectively solve the above problems.Because users’ behavior is predictable,if content deployment can take advantage of this feature,it can not only reduce transmission delay to improve user experience,but also reduce the energy consumption of the system,which is important for green networks.Therefore,how to predict users’ behavior and how to deploy and deliver requested content has become a very hot research area.A cooperative multicast proactive caching scheme based on adversarial automatic coding(AAE)is proposed in this thesis.This scheme achieves cooperative of small base stations and multicast transmission.In this scheme,firstly users are divided into different groups based on their characteristic.And then the content that the group may request will be predicated by using AAE.To reduce the redundancy of cached contents,the Ant Colony algorithm is used to pre-deploy the predicted contents to each small base station.In the content distribution phase,if a user requested a content with high popularity,the content will be proactively cached in a multicast manner to other users in this group who have not requested the content,otherwise it is distributed in a normal manner.Considering the users’ movement scenario,this thesis proposes a multicast and encoder collaborative Caching scheme based on adversarial autoencoders.Firstly,by using VAE prediction model and with users’ the historical data we can predict the users’ location.Then an algorithm which combining Genetic Algorithm with Artificial Bee Colony is proposed to solve the optimization problem of the relocation of different contend.Since multicast techniques can improve the transmission efficiency of the system and reduce the energy consumption,but each user may request different content.In order to have more chance of multicast transmission,in the content distribution phase,XOR encoding is used before data is transmitted.Through theoretical analysis and simulation,we verify that the proposed edge caching scheme for the two scenarios can effectively reduce the time delay and energy cost of the system.