Research On Resource Allocation Mechanisms With Qos Guarantee In Cooperative Wireless Networks

Wireless network is reforming our daily life,but the limited resource of a sin-gle wireless device,especially the communication resource,has severely hampered the enhancement of wireless network's quality of service.Thus,cooperative wireless net-works,which can improve the performance bound of a single device or even the w-hole wireless network by the means of cooperation,have gained tremendous attentions.There are two primary cooperative techniques,i.e.cooperative communication tech-nique and cognitive radio technique.Cooperative communication is an energy coop-eration strategy,under which relay node will consume energy to assist source node to transmit message,so source node can gain spatial diversity and gets its communication quality promoted.Cognitive radio is a spectrum cooperation strategy.In this strat-egy,primary user who has idle licensed spectrum permits secondary users who need spectrum desperately to access the spectrum dynamically so that spectrum utilization is improved.Cooperative wireless networks subvert the traditional wireless networks paradigm,but bring new challenges for resource allocation mechanism design as well.Even though there exist many resource allocation mechanisms which aim at improving the performance of a single communication link,the research on guaranteing the quality of service of the whole wireless network is still insufficient.Therefore,this thesis main-ly investigates the techniques of designing resource allocation mechanisms under the energy cooperation strategy and the spectrum cooperation strategy to ensure the quality of service of the whole cooperative wireless network,such as interference,capacity and social welfare.The primary contents and contributions of this thesis can be summarized as follows:1.Serious network interference will led to data retransmission and communication collision,so interference is a critical performance criterion for wireless network.Under the cooperative communication mode,although source node can exploit relay node's spatial diversity to improve its transmission quality,the interference caused by relay node can not be ignored as well.Therefore,we first consider a cooperative topology control problem whose target is constructing low interfer-ence network topology with cooperative communication technique.An interfer-ence model which is suit for both the traditional communication mode and the cooperative communication mode is proposed.Based on this model,we then formalise two problems:one tries to construct a min-max interference topology that is strongly connected and the other tries to build a min-average interference topology that is strongly connected.For the former problem,we design an opti-mal algorithm called DBS which is based on the binary search skill.Due to its NP-hardness,we proposed a distributed algorithm called DBCT whose approx-imate factor is 4 for the second problem.Finally,the simulation results demon-strate that DBS and DBCT can reduce network interference efficiently,and also decrease energy consumption by about 30%.2.In cognitive radio technique,spectrum auction is generally used to stimulate pri-mary users to adopt the spectrum cooperation strategy.To provide sufficient flex-ibility,auctions are periodically conducted in a round-by-round manner.In ad-dition,a critical property called social welfare is generally considered so that the auctioneer can allocate the spectrum resource to those secondary users who need spectrum desperately.However,previous studies only focus on social welfare op-timization in a single auction round,and rarely consider how to guarantee the so-cial welfare performance of multiple auction round from a long-term perspective.Therefore,from a long-term perspective,this thesis interstage the techniques for social welfare maximization in periodic spectrum auction under the budget con-straints of secondary users.Based on the primal-dual skill,a periodic spectrum auction framework called PSA and a tailored one round spectrum auction scheme called ORSA are proposed.We prove that ORSA has many economic properties,e.g.individual rationality,truthfulness and computational efficiency.With the dual fitting technique,we also prove that if PSA runs ORSA in each round,then our mechanism can both maximizes the social welfare in each single round ap-proximately and ensures only a small loss of approximation ratio from a long-term perspective.Extensive simulations show that PSA outperforms previous works on various metrics significantly.3.To ensure that spectrum resource can be utilized efficiently in secondary network-s,we study a joint channel and sink assignment problem thoroughly under the context of data collection in single-hop cognitive radio networks.Specifically,we study the influence of channel switch delay on network performance careful-ly.Then,to maximize the minimum capacity of all secondary sensors,we first resort to some elaborated mathematical skills to formulate the joint channel and sink assignment problem as a mixed integer linear program.Combined with lin-ear relaxation skill and greedy strategy,we give two heuristic algorithms.At last,we conduct extensive experiment to demonstrate the efficiency of our algorithms.4.Cooperative cognitive radio networks(CCRN)is a promising paradigm which fuse cooperative communication and cognitive radio perfectly.To stimulate both primary users and relay nodes to participate in CCRN,we discuss a auction mech-anism problem in a static CCRN scenario which has several primary users and multiple SU groups.Here,a SU group is a set of secondary users that are in-terested in sharing a common secondary relay.We first model our problem as a hierarchical auction,based on which we propose a truthful auction mechanism,i.e.TERA,and a fair auction mechanism named UERA.Specifically,TERA has a critical property known as truthfulness,whereas UERA is only an approximate truthful mechanism but can ensure fairness in a SU group.Those two mechanisms also have other critical properties like truthful,individual rationality,budget bal-ance,supply limits and computational efficiency.Moreover,we prove that both TERA and UERA can gain near-optimal revenue theoretically.According to our simulation results,the proposed mechanisms can provide enough capacity gain for secondary users,and improve the utility of PUs and relay nodes significantly up to 125%and 151%respectively.