Game Based Flow Rate Control for Networks

Traditional rate control protocols for networks such as TCP require cooperation between traffic sources to achieve optimal network performance. However, traffic sources may not be cooperative due to a widely heterogeneous demand of end-users. We propose a flow rate control framework using noncooperative game theory. The scheme is based on the idea of the Stackelberg solution from noncooperative game theory. The network as a leader designs a pricing mechanism for network bandwidth that attempts to drive the users' flow to the social optimal solution. Each user as a follower is charged by network, and chooses a willingness-to-pay to maximize his own net profit. We prove that the rate control game admits a unique Stackelberg equilibrium point and the bandwidth allocation is efficient and fair