Research On Resource Management And Games In Heterogeneous Wireless Networks

A ubiquitous and heterogeneous wireless access environment is provided to mobility users with the occurence and deployment of multiple wireless networks access technologies. In such a heterogeneous wireless networks(HWNs), individual wireless networks technology owes its own properties and characteristics, including data transmission rate, signal coverage, network capacity, QoS, security guarantee and the support to mobility. Different wireless access networks coexist, overlay mutually, cooperate with each other, complement reciprocally, supplying better QoS guarantee to mobile users. On the one hand, multi-mode mobile terminals could flexibly access to more appropriate wireless networks according to the QoS requirement of service and user’s preference, and guarantee QoS of roaming user. On the other hand, network operators can improve utilization ratio of wireless networks resource and capture more profits by means of collaboration of different wireless access networks.The resource management in HWNs is to implement effective configuration of wireless networks resource with relevant decision algorithms, scheduling mechanisms and coordination measures either jointly or cooperatively, through integrating and converging diverse wireless networks access technologies. Its major research fields involve in load balancing, network selection, vertical handover, admission control, seamless roaming and QoS guarantee. In this thesis, we elaborate on load balancing problem in HWNs from the perspective of system scheme and decision algorithm, and dedicate to raise the performance and efficiency of radio resource. In the process of researching, we expand research ideas via introducing diverse optimization theory methods, and resolve several problems in typical HWNs scenarios. The proposed schemes and algorithms contain both integrated and distributed management approaches, and simultaneously we take into account mobile users benefit, operator profits and community income. Combining theoretical demonstration and simulation experiments, we analyze and validate the performance of related algorithms. The proposed schemes hold certain reference value for the implementation and deployment of HWNs converging technologies.Specifically, the key contributions in the dissertation can be described as follows:(1) A radio resource management scheme of HWNs is presented to achieve user suboptimal max-min bandwidth fairness by association control in terms of tight coupling HWNs. In the HWNs environment, uneven traffic distribution among different wireless access networks results in the distinction of user network resource. Firstly we extend legacy IEEE802.21protocol, which can supply updated information of network status and user to radio management entity for central decision. A load balancing scheme is proposed to achieve user max-min bandwidth fairness by association control mechanism. The max-min bandwidth fairness allocation in HWNs is NP-hard problem, so an approximation algorithm of polynomial-time is designed to obtain suboptimal result. Final simulation demonstrates the designed algorithm’s performance in user bandwidth fairness and whole system efficiency.(2) With the aim at resolving load balancing of loose coupling HWNs, a user-network association algorithm based on population games is proposed to obtain user optimal access network and the throughput maximization of whole HWNs. Load balancing is an important issue in HWNs. We propose a user-network association algorithm based on population games to solve the problem. Firstly the user-network association is modeled as a population games which satisfies the requirement of potential games relying on the user utility function in access network. By means of replicator dynamics, we demonstrate that evolutionary result eventually converges to Nash Equilibrium. This makes sure each user connecting to a network with the maximum payoff. Moreover, it is proved that the throughput of whole HWNS can be maximized after reaching Nash Equilibrium, which ensures the efficiency of Nash Equilibrium. Finally, a user-network association algorithm is presented with the principle of replicator dynamics. Simulation shows that Nash Equilibrium point is achieved and the results of theoretical analysis are verified.(3) A resource allocation scheme in the independent HWNs is addressed to satisfy optimal utility of both mobile users and operators based on multi-leader multi-follower Stackelberg games model, achieving optimal bandwidth strategies of mobility users and pricing ones of network operators. Firstly we design user utility function with payoff and cost, and demonstrate that the utility function obey concave function condition for two kinds of service after defining operator price strategies, which assures the existence of Nash Equilibrium point. A distributed iterative algorithm is derived to obtain optimal bandwidth allocation and pricing strategies for mobility uses and network operators respectively. Each user modifies its bandwidth proportional to the gradient of its utility function with respect to its bandwidth, and the algorithm eventually converges to Nash Equilibrium point. Afterwards, network operators iteratively adjust respective price by marginal utility, and capture optimal pricing strategies at Nash Equilibrium point. Simulation experiments model iterative process and acquire the solution of Subgame Perfect Nash Equilibrium.