The Game On Complex Networks And Its Application For Resource Management In Communication Networks

The three great discoveries of network science (small world phenomenon, free scale behavior, superfamily characteristics) greatly stimulated the enthusiasm of scientists from every field on the research of complex networks. The works of complex networks are permeating many different fields, such as mathematics, physics, life science, economics, and engineering. The study of real networks from the perspective of complex networks is receiving more and more attention. Owing to the advantage on dealing with the conflict and cooperation on real networks, especially the advantage on the study of the emergence of group cooperation behaviors as a result of individual adaptability, the game on complex networks can provide a general theory for the macroscopic description of nature, social and life. Therefore, it has very broad research prospects. We first studied the game on complex networks and then applied it on the resource management of communication networks. The main work is as follows.Firstly, we explored the spatial evolution behavior of complex networks. By analyzing the disadvantaged long-range connection behavior in weighted networks, we constructed a simple field theory in networks. According to the field theory in network, topological growth mechanism and preferential attachment mechanism, a complex network space-time evolvement model was proposed. Based on interactive-force-driven network evolution dynamics, the model integrates the strength, topology, and distance and extends the BBV model from space. In order to present the networks generated by the model, a method to construct weighted networks on lattices was suggested. The networks generated by the model on lattices exhibit the statistical properties observed in several real-world systems, such as the spatial collectivization phenomenon, small world property, and scale-free behavior. All of this work can be recognized as a general starting point for the realistic spatial modeling of complex weighted networks.Based on the research of complex network topology, we studied the game on complex networks. Through analyzing the learning mechanism of individual on complex networks, the complex of the individual learning model in realistic networks was pointed out, and a Prisoner's Dilemma Game (PDG) Model based on Hybrid learning Mechanism (PDGM-HLM) was proposed. We also discussed the influence exerted on the learning mechanism by the network structure and the value of temptation to betray. The simulation experiments provided the results as follows. When the value of temptation to betray is little, the learning mechanism of imitating the strategy of neighbors can enhance the cooperation of networks. While, the learning mechanism based on historical memory is litter affected by the value of temptation to betray. The HLM can make the cooperative individuals survive when the value of temptation to betray is great. Meanwhile, the structure of networks has a certain effect on the cooperation level.Then, we studied the co-evolution of weighted networks and game dynamics, and a model for the co-evolution of weighted networks structure and game dynamics was proposed. By employing the mean field approximate theory and simulations experiments, we explored the co-evolution behavior on weighted networks. The simulation experiments provided that the co-evolution of network structure and game dynamics can be reappeared by the model. The wealth and strength have same distribution character. Meanwhile, the networks generated by the model exhibit the statistical properties observed in several real-world system, such as small world property and scale-free behavior.Based on the above three aspects, the application of the game on complex networks in communication network resource management was discussed. Under the application framework, the flow control in internet and the distributed power allocation in wireless cooperative networks were studied.Through analyzing internet resources management, the game on complex networks was applied on the flow control of internet. By employing the Stackelberg network game, a general model for network flow control was proposed. Under this model, two flow control algorithms based on differentiated pricing scheme and uniform pricing scheme were proposed respectively. Simulation results show that the proposed algorithms can enhance the revenue of Network Service Provider (NSP) and the network performance without significantly affecting user satisfaction, compared with the Asymptotical Solutions of Tamer Basar (AS-TB).Through analyzing the distributed power allocation in wireless cooperative networks and its frameworks, the game on complex networks was applied on the distributed power control in wireless cooperative networks. A symmetric system model for power allocation between cooperative nodes was proposed. Referring to the evolutionary game algorithm (EGA), a distributed power control algorithm based on the EGA was proposed. The algorithm can guide the nodes'cooperative behavior and enhance the resource utilization, compared with the directed transmission.