Mobile Data Offloading Research Based On Incomplete Information Game Theory

With the continuous development of mobile devices,people's demand for multimedia services has also increased dramatically,which lead a rapid increase in data traffic requests from cellular networks.As a result,the traditional operator's network architecture is difficult to solve the contradiction between the increase in user demand and the lack of operator traffic.However,the introduction of third-party agent for mobile data offloading not only significantly reduces the load on the cellular network operator but also improves user satisfaction.Due to the isolation of information in the actual environment,operators and users cannot effectively obtain global information.Therefore,the research in this paper is mainly based on the multi-operator mobile data offloading under incomplete information.First of all,we analyse the mobile data offloading problem by game theory,and describe its scenario by classification.According to the problem of existing works,we propose the cooperative and non-cooperative mobile data offloading model under incomplete information.In the first model,we discuss the problem of maximizing the overall interests of users and minimizing the overall interests of operators.The purpose in this situation is to maximize overall social welfare.Later,in the non-cooperative offloading model,we perform a more realistic analysis of users and operators,which pursues the maximum interests of users and operators.Secondly,for the research of mobile data offloading based on cooperation,we aim to maximize overall social welfare by combining the maximization of user satisfaction and the minimization of operator interest.For this problem,we solve it by introducing a third-party agent,which can use the water-filling algorithm to allocate the traffic for the user.In the process of interaction between the set of the users,operators and the agents,the problem of maximizing social welfare is solved through a distributed iteration algorithm.In addition,the performance of the proposed method is verified in different situations.The proposed method can converge quickly and the result of distributed method is in line with the result of centralized method.Finally,when rational users and operators offloading under the condition of incomplete information,they all pursue the maximum self-benefits.Therefore,we model the problem as a multi-user multi-operator non-cooperative game.Then,we propose a traffic allocation algorithm based on marginal utility,which can allocate the traffic and iteratively adjust the transaction price between users and operators through the bidding model.It ensures the efficiency of the model's convergence.Meanwhile,we analyse the game structure by game theory,and prove that there exists a Nash equilibrium.Besides,we give a corollary about price and verify the existence of Nash equilibrium by experiments.