The Research Of Network Resource Allocation Based On Stackelberg Game

In recent years, with the network traffic growing exponentially, the demand of a variety of applications increases rapidly. In particular, the real-time multimedia applications that require strict QoS need more network resources, thus an efficient resources allocation scheme is important. Fortunately, Game Theory has been widely applied to management, ecology and industrial control as the tools of analysis and solution to the conflict and cooperation. The current research hotspot is the introduction of market mechanisms in order to manage network resources.This paper introduces master/slave idea in game theory to network traffic control to optimize the performance with multi-objectives. Stackelberg Game Theory is built based on studying Multi-local-world model. The existence of Nash equilibrium is proven. The optimal price and flow rate under the equilibrium state which can describe the optimal network state quantificationally are given. The effect of network topology on network equilibrium is analyzed and network evolutive characteristic caused by the network topology changing is studied on this basis. Matlab and NS2 software are used to simulate and show that if the system resource is sufficient, the more nodes there are in the network, the more benefit the system and users will get; if the system resource is poor, the new nodes will be prevent by the nodes located in the net to access in the network. More resources provided by network system for users are needed. It may be one of the reasons why the network grows quickly and constantly.Based on the Stackelberg Game model, this paper introduces the incentive strategy and proposes a network adjustment algorithm which can forces network from general state to optimal state. The team-optimal solution ensures the stability of network in the optimal state. Simulations test the effectiveness of the algorithm.Finally the entire work is reviewed and some suggestions are given for the research in future.