Resource Management Technology For NOMA-Enabled Dense Wireless Networks

The future wireless networks aim to support Internet of everything and ubiquitous communications.Improving network capacity and connectivity is a powerful approach to achieve this purpose.The academia and industry have proven that network densification can significantly enhance network capacity,and non-orthogonal multiple access(NOMA) can breakthrough the bottleneck of connectivity on limited spectrum resources.As a consequence,NOMA-enabled dense wireless networks are regarded as important development trends for future wireless networks.Compared with traditional orthogonal multiple access,NOMA breaks the orthogonal constraint of resource assignment and allows multiple users simultaneously to share the same resource,which fundamentally changes the principle of resource assignment.In dense wireless networks,the user-base station(BS)association becomes multi-point association from single-point association,and the network service transits from BS-centric to user-centric,which lead to the corresponding adjustment and improvement of resource management techniques.Therefore,how to design efficient resource management techniques such as user-BS association,spectrum allocation,and power control to release the potential of dense wireless networks and NOMA on network capacity and connectivity is crucial and worthy of investigation.In NOMA-enabled dense wireless networks,non-orthogonal resource reuse introduces co-channel interference.The strength of co-channel interference is determined by not only signal power levels but also the decoding order of users,which complicates the interference relationship between users.Hence,traditional resource allocation and power control algorithms are difficult to coordinate complicated interference in NOMA enabled dense wireless networks.Furthermore,the number of users and that of access points are very large,and the resources on frequency/time/space/power domains coexist,which greatly increase the problem dimension and the algorithm complexity for resource management.Hence,it is very challenging to realize efficient multi-domain resource management while satisfying resource constraints and user requirements.According to the above discussion and challenges,we investigate resource management techniques in NOMA-enabled dense wireless networks,aiming maximizing network capacity and connectivity.From two aspects of network-dense deployment and user-dense access,resource management algorithms are designed for different kinds of dense networking means.Proposed algorithms can achieve efficient matching between user requirements and resources to improve system spectrum efficiency and the number of connected users.The contribution of this dissertation is listed as follows:1.We propose a joint optimization algorithm of BS clustering and power control to maximize system spectrum efficiency for NOMA-enabled cellular networks,which alleviates the degradation of network capacity due to network over-densification.This work first characterizes interference in BS clusters caused by non-orthogonal spectrum reuse and analyzes its impact on joint transmission rate from multiple BSs.Then,we derive a tight lower-bound to system spectrum efficiency,which is exploited to optimize power control and BS clustering individually.Based on this,The alternating optimization method is adopted to design a joint optimization algorithm of BS clustering and power control which maximizes system spectrum efficiency while meeting the access requirements of all users.Theoretical analysis shows that the proposed algorithm can obtain a series of non-decreasing values of system spectrum efficiency and have asymptotic convergence.Finally,simulation results validate the correctness of theoretical analysis and the effectiveness of algorithm design and show that the proposed algorithm can harvest the gain of dense deployment of BSs to efficiently mitigate the degradation of network capacity due to network over-densification.2.We propose a unified management algorithm of mode selection and resource allocation to maximize system spectrum efficiency,which restrains strong interference between cellular users and dense device-to-device(D2D) pairs in NOMA wireless networks.Specifically,this work designs the interlay mode as a new resource sharing mode for cellular users and D2D pairs that utilizes the advantage of power-domain multiplexing of NOMA to cancel strong interference between cellular users and D2D pairs.Then,maximum weight clique in graph theory is used to design a resource management algorithm,which uniformly optimizes mode selection,subcarrier assignment,and power control.The proposed algorithm can obtain the optimal solution of mode selection and subcarrier assignment and exploits the special structure of the graph model to prune the search space to reduce computational complexity.Finally,simulation results validate the effectiveness of proposed algorithm and efficiency of graph model and show that the interlay mode and proposed algorithm can effectively coordinate cross-tier interference between cellular users and D2D pairs to avoid the decline of network capacity caused by a large number of D2D pairs.3.We propose a dynamic resource management algorithm to maximize the average delay of users,which relieves the congestion issue when a massive number of users upload traffic.This work focuses on an issue that the differentiated uploading delay between users could cause the variation of co-channel interference during the traffic uploading process.For this issue,we reveal the interaction between differentiated uploading delay and co-channel interference between NOMA users.Theoretical analysis shows that if the uploading completion order of users is inconsistent with their decoding order,reducing co-channel interference cannot decrease uploading delays of both NOMA users.Based on this,the optimal power allocation algorithm is designed which dynamically adjusts the transmit power of users according to the variation of the difference between uploading delays to minimize the average delay of a pair of NOMA users.Then,the relaxation method of optimization theory is exploited to design a joint uploading decision and resource allocation algorithm.Finally,simulation results validate the correctness of theoretical derivation and the effectiveness of the proposed algorithm and show that the proposed algorithm can efficiently reduce the average delay of users and breakthrough the bottleneck of traffic uploading capacity.