点击次数：

论文类型：期刊论文

发表时间：2013-12-01

发表刊物：APPLIED INTELLIGENCE

收录刊物：SCIE、EI、Scopus

卷号：39

期号：4,SI

页面范围：692-704

ISSN号：0924-669X

关键字：Ad Hoc network; Channel assignment; Dynamic game with perfect and complete information; Backward induction; Nash equilibrium

摘要：The nodes in Ad Hoc networks compete for channels when communicating, with the features of no center and self-organization. In traditional channel assignment strategy of MAC layer, nodes do not consider the demands to channel resources of other nodes, which hinders improving the network performance. Practically, in the network based on competitive MAC protocol, each node tries to maximize its payoff, while this interferes with the behavior of other nodes at the same time. Game theory is an effective tool to solve problems of distributed resources, which can be used effectively in channel assignment. In this paper, we propose a new protocol, namely, DGPCI-DCA (Dynamic Game with Perfect and Complete Information based Dynamic Channel Assignment). When all the nodes are rational and greedy, each node selects channels dynamically by backward induction according to strategies of other nodes, thus Nash equilibrium can finally be achieved. Experiments show that the network performance is effectively improved, i.e., the throughput and saturation throughput can be increased, and the packet loss rate and network delay can be reduced.