Research And Simulated Implementation Of Multi-Agent Cooperative Mechanism For Mobile Social Cloud

With the rapid development of cloud computing,mobile devices and mobile applications,mobile social cloud computing has become a very promising technology.Mobile social cloud is a social interaction mobile platform,which is based on collaboration as the core,resource sharing as the main purpose,and there are social relationships among members of the cloud.As a result,method can be used to combine the resources in the cloud,enabling users to accomplish tasks that are difficult to accomplish with a single device through social collaboration,while the resources in the mobile cloud can be used effectively.According to the social characteristics of mobile social cloud,this thesis proposes a collaborative strategy of Q reinforcement learning,which provides a solution for mobile social cloud resource cooperation based on the multi-agent theory and Markov process theory.The task is to study how to allocate resources to users of the cloud as efficiently as possible and to simulate the resources collaboration of mobile social cloud.Since topological relationships in social networks are often based on real world relationships,this thesis enables users to share resources through learning and collaboration in a social network environment by using the characteristics of these social relationships.In this thesis,an improved Q-learning algorithm based on the characteristics of social network is proposed as users' cooperation strategy,and a request-response strategy is proposed for the request conflict,and a simulation experiment is carried out.Finally,the algorithm is compared with algorithms presented in the literature about the colaboration effectiveness.The comparison shows:the collaborative method proposed in this thesis can improve the utilization of resources.This thesis first introduces the research background,research status,research tasks and significance,and then start with the related technology,the thesis introduces the multi-agent reinforcement learning theory and social network analysis technology.Then under this model,a mobile social cloud simulation model based on small-world network is established,and simulation experiments are carried out,and the comparison of various cooperation methods is made.Finally,the advantages and disadvantages of this study are summarized and the future work is prospected.