Research On Resource Allocation Technology Of Virtual Radio Access Network

With the emergence of the fifth generation(5G)network,wireless net-work virtualization has become more important,and users' requirements for high rate,low latency and ultra-reliable communication are becoming stronger.Network slicing technology is essential to meet the QoS requirements of users in different industries.Based on the background of 5G virtual radio access network(RAN),this thesis studies the optimization technology of slicing re-sources in different traffic scenarios,eMBB and URLLC.Firstly,from the per-spective of multi-service slicing,according to different service types,the corre-sponding network slicing management scheme is designed.Then optimize the system indicators such as spectral efficiency,energy consumption,users' util-ities,and operator's revenue,which involves the utility equilibrium between infrastructure provider(InP),mobile virtual network operator(MVNO)and users,as well as the inter-slice and intra-slice resource optimizations.Finally,MEC,as one of the key technologies to promote 5G,can provides cloud com-puting function in the edge of the RAN.Therefore,we design an optimization algorithm for joint communication and computing resources,which can opti-mize the utilization rate of system resources and reduce the delay of end users.Regarding the abstract of this article,we introduce it from the following four aspects:1.Multi-service slicing resource optimization scheme of joint power and subcarrier in virtual RAN:In virtual RAN,this thesis studies the op-timization of wireless virtual network slicing resources.Multi-services' users accessing the same network causes interference among different services,and the spectral efficiency in RAN cannot be guaranteed.At first,the isolation be-tween different services can be guaranteed by introducing RAN slicing,i.e.,different services access different virtual RAN,in which the virtualized RAN is divided into eMBB slicing and URLLC slicing.Then the proposed solution is modeled as an optimization problem,in which the constraint is to ensure the isolation inter-slice and meet the requirements of users accessing two slices,and the goal is to maximize the spectral efficiency of the system.This problem is a mixed integer programming problem.By transforming and relaxing integer variables,the original problem is approximated as a convex optimization prob-lem.Finally,by means of dual variables,the objective function and constraints are coupled into an augmented Lagrangian function,and the optimal solution of this function is the upper bound of the original problem.In addition,a resource allocation algorithm is proposed,which combines PHR and branch and bound method to provided the optimal allocation strategy of network slicing,i.e.,the approximate optimal solution of the original problem.Simulation results show that compared with APSO,EPA and ESA,the proposed resource allocation al-gorithm can significantly improve the spectral efficiency of the system and the reliability of URLLC.Furthermore,with the users' requirements change,this thesis analyzes the spectral efficiency of the proposed algorithm and get better spectral efficiency performance.2.Multi-tenant heterogeneous network slicing resource optimization scheme based on MPEC:In multi-tenant heterogeneous RAN,network slicing technology permits multiple tenants,i.e.,MVNOs,to share virtual resources.However,each tenant only considers the individual slice utility,which leads to the unfair resource allocation among tenants.To achieve the aim that the InP can fairly allocate virtual resources to tenants,this thesis proposes a two-layer re-source allocation architecture in heterogeneous RAN and formulate it as a math-ematical program with equilibrium constraints(MPEC).The existence of the solution in the lower layer is proved via the properties of the quasi-variational problem(QVI),indicating that the MPEC is solvable.Combining the two-layer architecture and successive convex approximation(SCA)method,this thesis has proposed a fair algorithm,which provides fair resource allocation strate-gies for the InP.Compared with the existing static slicing(SS)and social opti-mal(SO)methods,the simulation results confirm that the proposed algorithm weighs the utilities of the total network system and each tenant.In addition,the gap between the proposed method and the SO regarding their utilities is within 5%,which outperforms SS.3.Resource allocation equilibrium scheme inter-slice and intra-slice in cognitive networks:The cognitive network has become a promising method to solve the spectrum resources shortage problem.Especially for the optimiza-tion of network slicing resources in the cognitive RAN,this thesis investigates the profit of the MVNO and the utilities of secondary users(SUs).In cognitive networks,the PUs' current problem is how MVNO virtually slices the redun-dant resources and efficiently allocates them to SUs.Since the MVNO and SUs are selfish and the game between the MVNO and SUs is difficult to reach equi-librium,this thesis consider modeling this scheme as a Stackelberg game and transforming it into MPEC problem to solve.Leveraging Karush-Kuhn-Tucker(KKT)condition,we can obtain a single-level optimization problem,and then prove that the problem is a convex optimization problem.The simulation results show that the proposed method is superior to the non-cooperative game.While guaranteeing the QoS of primary users(PUs)and SUs,the proposed method can balance the profit of the MVNO and the utilities of SUs.4.Dynamic network slicing resource optimization MEC scheme for long-term stability in joint communication and edge computing networks:In the joint communication and edge computing networks,this thesis consid-ers two kinds of slices,one is a traditional communication slicing with eMBB requirements.The other is the URLLC slicing with edge servers.For edge computing slice,this paper designs a variety of alternative computing methods,such as local computing,fog computing and edge computing,according to the amount of computing traffic arriving.When the arriving computing traffic can-not be completed in one time slot,this paper offloads the computing traffic to the fog node or edge node.There are two problems to be solved.The first is how to get the optimized unloading strategy that can meet the low delay of users and minimize the energy consumption of users.Secondly,it is the problem of slicing resource allocation between communication resources with computing task unloading and traditional communication resources.Therefore,how to share RAN communication spectrum resources is a key factor.This thesis use Lyapunov optimization to study the dynamic resource optimization of multi-services based on network slicing technology,in which the services include eMBB traffic with high rate requirement and URLLC traffic with low delay re-quirement.This thesis maps two kinds of slicing service processes in long-term network into a Lyapunov queue,solves different slicing service requirements by convex optimization method,and minimizes the energy consumption of the system.Simulation results show that the proposed scheme can guarantee the long-term stability of the network system.The backlog of queues is conver-gent.More importantly,we show the stability of the two slices meeting the real-time requirements of users.