A Game Theoretic Framework for Quality of Experience Enhancement in Dense Stadia

In crowded venues, such as sports stadia, maintaining an acceptable network quality of experience (QoE) is hard to achieve. Installing small cells, distributed antenna systems or high-density WiFi in every stadium is too expensive for mobile network operators. Hence, we propose a novel distributed low-cost solution based on user coordination to improve the average QoE when the network capacity cannot be enhanced. Specifically, fans take turns in disabling their cellular connectivity, such that the connected users utilize the relaxed network to obtain then share common match data with the disconnected users via Peer-to-Peer (P2P) connectivity. To eliminate a free-riding behavior, a limited punishment strategy in a large repeated game is proposed and shown to yield an approximate subgame-perfect Nash equilibrium. In addition, we model human irrationality as game noise incorporated into the proposed equilibria. A proposed application-oriented QoE model is first obtained via SimuLTE, an extension of OMNET++, then used in MATLAB simulations to verify the proposed solution. The results show tangible gains realized by the proposed solution under realistic scenarios and parameters.