Research On Distributed Resource Allocation Algorithms For Next Generation Of Wireless Communication Networks

With massive users and devices pouring into wireless networks,network densification turns out to be an inevitable trend of the development of future wireless networks,where resource allocation algorithms with distributed and self-orgnized feature become an inherent need.Meanwhile,the vigorously growing mobile Internet services also placed new requirements on network service abilities.On the one hand,as the mobile payment penetrates into every corner of our life,unprecedented attention has been triggered to the informa-tion security,raising new demands for future network capabilities.On the other hand,the prevalence of video services has also announced an urgent demand for efficient video caching and delivery,which leads to a new network architec-ture where the computing and caching resources are sinked down to the edge.Researches on distributed resource management to handle challenges brought by the new demand as well as the new architecture have become an essential part of the network development.Recent years have witnessed a growing interest in physical layer security as an effective tool to improve the information security of wireless networks.However,most of the works are carried out in relay and cognitive scenarios,or investigate the signal processing in device-to-device(D2D)networks with-out considering the feature of complex and large scale networks.This limit-s the performance of physical layer security techniques in future ultra-dense networks(UDNs).On the other hand,the expansion of resource dimensions in wireless edge networks make it more challenging for joint resource man-agement,since traditional models are not capable of analyzing the multi-role problems as well as guaranteeing stable solutions.Hence,the distributed col-laboration mechanism with stable structure among different kinds of resource nodes has become an essential problem.Therefore,this thesis mainly focus on the distributed resource allocation in the field of application of physical layer security in UDNs,and stable cooperation schemes in wireless edge networks.The main contents and contributions of this thesis are summarized as fol-lows:1.Considering the inter-layer interaction in UDNs,a cooperative regime between two layers in UDNs for full buffer services is proposed to enhance the security transmission,where the small cell base stations(SBSs)act as the friendly jammer,assisting the information security of the macro layer.A re-source along with the monetary reward mechanism is designed for SBSs to boost the cooperation,after which the two-way selection problem between S-BSs and macro user equipments(MUEs)is modeled as a coalition formation game with non-transferable utilities by regarding the MUE and its jamming S-BSs as one coalition.Finally,the merge-and-split algorithm is exploited to find the Nash stable structure in a distributed way,the outcome of which remark-ably enhances the secrecy capacity of MUEs without deteriorating the SBSs'performances.2.Considering the intra-layer interaction in UDNs,distributed frequen-cy management algorithms for burst services are studied to efficiently enhance the macro layer security in the whole working bandwidth.The cross-tier in-terference,treated as evil by traditional wisdom,is employed to confuse the malicious node as well as to tackle the conflict between secrecy and tradition-al performances through the prudent spectrum allocation among small cells.Three games are designed to progressively give insight into the frequency as-signment problems under increasingly strict scenes.The secrecy performance is considered at first by leveraging the state-based potential game,which guar-antees a budget-balanced utility design.The existence of a recurrent state equi-librium point is proved,and that it is able to maximize the total safety transmis-sion probability of all subchannels.A step forward,the service delay of SBSs is taken into consideration.To this end,an exact potential game,in which the equilibrium always exists,is built to help SBSs strike a balance between altruistically helping the macro users and selfishly keeping their own perfor-mances.Furthermore,a fully distributed non-cooperative game that requires no exchange among SBSs is put forward.The proposed scheme may work well even when the backhaul is limited or even unavailable since all SBSs only depend on the observation of their own instantaneous performances.3.For resource allocation in Content Centric Networks(CCNs),the coop-eration mechanism in hierarchical in-network caching framework is taken into account,where SBSs and macro base station(MBS)take the role of the CC-N router,handling the Interest packets released by users.Presuming that the SBSs in a hot spot collaborate to maximize their overall revenue by network selection together with the bandwidth allocation,an optimization problem is formulated with the objective of maximizing the overall net earning as well as users' delay constraint.Subsequently,the payoff is divided among SBSs via a bankruptcy game,and the solution is found employing the Nucleolus fairness concept,which guarantees the stable cooperation.A decentralized Nucleolus calculation algorithm is adopted here so that the network can work in a dis-tributed manner.4.For joint configuration of multi-resources in mobile edge computing networks(MEC),schemes based on network virtualization to effectively man-age the computing and communication resources are proposed,in which mo-bile virtual network operators(MVNOs)create virtual slices to serve the re-quests issued by their subscribed users.It is first assumed that the virtual duplications of every content are created for each user,so as to decouple the many-to-many relationship between the contents and users,followed by the op-eration of combining the virtual duplications and the virtual slices as resource pairs.Consequently,the four-sided matching among users,contents,fog n-odes and MVNOs are compressed into a three-sided matching among users,resource pairs and fog nodes.The above problem is formulated as a three-sided assignment game exploiting the supplier-firm-buyer(SFB)game model,which gives the optimal solution of matchings among the contents,computa-tion nodes,MVNOs,and users.Moreover,the existence of the non-empty core of such game is proved,indicating that the proposed framework is stable.In order to obtain the simple practical solution,a distributed suboptimal algorith-m of reduced version of three-sided matching with size and cyclic preference(R-TMSC)is adopted.Furthermore,a greedy strategy is proposed to improve the convergence speed as well as performance of the R-TMSC scheme.