A Game Theory Approach For AP Association And Resource Allocation

In recent years, with the increase in the number of mobile users and the rapid development of built-in wireless network mobile portable devices(such as flat, mobile phones, notebooks, etc.), people have more and more interests on the Internet, and hope to access the Internet anytime, anywhere.In order to meet people's needs, wireless local network(WLANs) based on the IEEE 802.11 has been developed rapidly, due to the characteristics with low cost, easy expansion, flexibility and other characteristics compared to other wireless networks. However, due to the limited wireless resources in the network and the rapid increase in the number of mobile users, the network performance of wireless LANs deployed by dense APs is greatly challenged.In the wireless LAN with Aps densely deployed, many user terminal nodes are covered by multiple APs. By default, each user chooses the AP with the largest signal strength to associate. This RSSI-based association method can lead to the imbalance of AP load and network performance degradation. Especially after introducing the multi rate, the DCF mechanism adopted by IEEE 802.11 will lead to the IEEE 802.11 performance anomaly, and affect the overall throughput of the whole network seriously. In order to provide a better network service experience for mobile users in the network, optimizing the association mechanism between AP and users has become a hot spot of network research in recent years. There exists a strong competitive relationship between users and the goals of users and users, as well as users and APs are interactive. Thus,how to balance user fairness and network throughput is the urgent need to solve in AP Association problem.Matching game as a model of game theory, it is not only a mathematical tool to deal with the competition relationship of participants,but also different from other game models. Matching game can consider different optimization goals of multi-party market participants, and allow the participants in the game have different income goals. Therefore, this paper introduces the matching game to solve the above problems. First, in order to mitigate the impact of IEEE 802.11 anomalies on network performance, we introduce a time-based bandwidth allocation mechanism to the AP association problem. Then, based on the time-based bandwidth allocation mechanism, we consider the AP association and bandwidth allocation problem jointly. Different from the existing research, the two-optimization goals of throughput and fairness are transformed into a single-objective optimization problem based on proportional fairness. In this paper, the matching game is introduced to model the association problem between user and AP, a matching algorithm is proposed to solve the association problem, and then, we propose a matching algorithm to solve the related problems. We make a brief description about the relationship between the throughput and fairness. Finally, we do the simulation of our algorithm on the MATLAB platform. Firstly, we will prove that the performance if bandwidth allocation mechanism based on the time fairness is better than that of mechanism based on the default DCF. Secondly, we describe the equilibrium relationship between throughput and fairness, and compare our proposed matching algorithm with FAME. Finally, we make an analysis on the performance of our matching algorithm and the load of AP.