Game Theory-based Wireless Network Resource Allocation For Virtual-real Fusion Services

With the progress of Virtual Reality(VR),augmented reality and mixed reality,as well as the development of wireless communication technology,the application demand for immersive virtual-real integration technology is rising rapidly.Based on the resources provided by wireless networks,virtual-real fusion services provide immersive user experiences by integrating physical and virtual spaces.The fifth-generation mobile communication systems make services closer to users by deploying more resources at the edge of networks.However,the construction of virtual-real fusion ecology and the service demands of users have brought a series of challenges to the use of limited wireless network multidimensional resources,including real-time computing,trusted sensing data sharing and high-quality user experiences.Meanwhile,the large number of smart devices in wireless networks and the increased number of interactions between entities in virtual-real fusion services increase the complexity of the traditional centralized resource allocation scheme.Game theory can provide distributed theoretical solutions to complex interactions between rational decision-makers,which makes it suitable to study the strategic interactions between individuals or groups that occupy limited resources in virtual-real fusion services.However,the complex interaction between users in virtual-real fusion services will also bring new challenges to constructing the game model and obtaining the game equilibrium.Therefore,the wireless network resource allocation for virtual-real fusion services is studied based on the game theory to optimize resource allocation and improve entities’ utilities.The main contributions are given as follows:Firstly,considering that network characteristics in virtual-real fusion service will affect the allocation of edge computing resources,the edge computing resource allocation scheme is studied under the influence of network effects and congestion effects.Based on the threestage Stackelberg game,the influence of network effects and congestion effects on the interaction strategies and utilities of different entities are studied,which is expected to improve the satisfaction of edge servers,virtual-real fusion service providers and users on computing resource allocation.The optimal solutions of the proposed three-stage Stackelberg game are obtained based on the backward induction method.Simulation results show that the utilities of entities increase with the increase of social coefficient among users under the influence of the network effect.Meanwhile,the congestion effect reduces the average utility of users by up to 49%.Compared with the baseline,the utilities of edge servers,virtual-real fusion service providers and users can be improved with the three-stage Stackelberg game approach.Secondly,to alleviate the heavy computing load of edge servers caused by a large number of rendering tasks in virtual-real fusion services,a collaborative computing solution for mobile devices based on blockchain and Coded Distributed Computing(CDC)is proposed.The virtual-real fusion service provider can offload large-scale matrix computing tasks to several mobile devices for distributed collaborative computing.A joint game model is formulated based on the coalition formation game and Stackelberg game,which can realize the selection of reliable computing workers and the computing resource allocation incentives.The simulation results show that the average reputation values of selected computing workers are less affected by the malicious behavior percentage of computing workers with the proposed joint game-based reliable distributed computing scheme.Compared with the traditional best-effort worker selection scheme,the utility of the service provider is maximally increased by 17%,and the average utility of computing workers is maximally increased by 14%.Thirdly,aiming at the problem of trusted sensing data resources sharing for virtual scene reconstruction in virtual-real fusion service,a sensing data sharing scheme based on blockchain and edge computing is studied to promote the sharing of sensing data.A consensus mechanism based on the proof of data storage and transmission is proposed.Edge servers can obtain consensus rewards from blockchain networks by contributing storage and transmission resources to the sensing data-sharing process.Then,the interaction of sensing data storage and transmission resources between edge servers and sensing devices is modeled as a multi-leader and multi-follower Stackelberg game.The alternating direction multiplier method is adopted to design the game equilibrium algorithm for the sensing data resource-sharing game.Simulation results show that the proposed algorithm enables both sensing devices and edge servers to obtain stable optimal strategies.Compared with the traditional edge servers collaboration scheme,the total profit of edge devices is maximally increased by 59%,and the total utility of data owners is maximally increased by 52%.Fourthly,the modeling and maximization of users’ Quality of Experience(QoE)in virtual-real fusion services is studied.A QoE model based on the virtual distance and network effect is proposed.To measure the virtual distance between users,a new metric called ‘meta-distance’ is modeled.Then,a joint resource allocation and virtual-real fusion service selection scheme is proposed to solve the QoE utility maximization problem.In the inner loop of the proposed scheme,a one-to-many matching game is adopted to allocate computing and spectrum resources.In the outer loop of the proposed scheme,the hedonic coalition formation game is formulated to select virtual-real fusion services.Simulation results show that compared with the baselines,when the total number of subchannels is 120,the average meta-distance of users is reduced by 22.1%,and the QoE utility of users is improved by 11.2%.In conclusion,this thesis studies the wireless network resource allocation for virtualreal fusion services based on game theory.The computing resource allocation gametheoretical approaches under the influence of network characteristics and the cooperation of mobile devices are studied to solve the problem of real-time computing.Then,the trusted sensing data resource sharing game-theoretical approach based on blockchain is studied to solve the trusted sensing data sharing problem for virtual scene reconstruction.Finally,the QoE optimization and multidimensional resource allocation game-theoretical approach are studied to satisfy users’ high-quality QoE.The utilities of entities in virtual-real fusion services are improved by optimizing wireless network resource allocation.The work of this thesis may guide wireless resource allocation and optimization for virtual-real fusion services in the future.