Game-based Traffic Selection Scheme In WiFi Offloading

With the updating of wireless and mobile communication technology,the gradual increase of fixed broadband penetration,and the rise of multimedia applications,mobile traffic is tremendous and growing rapidly.In order to alleviate the contradiction between transmission demand and network supply,researchers consider transferring mobile data into WiFi network to alleviate network load in hot areas,improve the transmission quality of mobile users,and drive new business models,which arouses wide interest of all parties.In response to the wave of storage,computing and network resources sinking to the wireless edge,previous studies have proposed to establish a cache-enabled offloading system in the WiFi access point's local area networks(LAN),which can provide a store-and-forward upload fashion for delay-tolerant data by means of the storage capacity of common devices locally,in order to further enhance the performance of WiFi offloading.However,compared with the previous network selection problem in WiFi offloading,users in the new scenario need not only to pay attention to how to select the network,but also to consider how to upload data,which is the traffic selection problem described in this paper.As a classical method to study resource competition among rational individuals,game theory is widely used in resource management of wireless networks.Among them,Network Congestion Game and Bottleneck Routing Game evaluate users' payoffs from different views,but whether they are applicable in the above new scenario remains to be analyzed.In addition,due to the diversity and complexity of the offloading mode,the difference between the theoretical performance and the actual performance of the algorithm is also worthy of attention.Based on the existing technology architecture and game model,this paper studies the traffic selection scheme under the above scenarios.Firstly,the research status is investigated and analyzed,and then the non-cooperative game model is established.Then,some equivalents with the existing game models are analyzed and proved.The existing algorithm is improved and its decision-making rules are analyzed.Finally,an approximate algorithm for traffic selection is proposed,and the performance is evaluated by numerical simulation.Compared with the improvement of existing algorithm,the proposed approximation algorithm effectively reduces the time cost of solving the problem,and the utility loss is relatively minor,achieving a good trade-off between performance and operational efficiency.In addition,in order to further evaluate the performance of traffic selection algorithm in a real network environment,this paper establishes a network experimental environment suitable for the system model,runs network experiments in indoor scenarios,and compares the results with numerical simulations.Finally,the scalability of the experimental environment is introduced through two cases.In view of the current situation of game-based research focusing on numerical simulation,the application of network experimental environment provides a reference for similar research on evaluation methods.