Research And Implementation Of Fine-grained Resource Slicing And Scheduling Algorithm For 6G Networking

With the continuous development of network technology,the requirements for networks are constantly increasing,and future 6G networks will form ubiquitous intelligent scenarios with global coverage and massive connections.Compared with 5G networks,services will be more granular according to services,and more indicators and functions will be generated,such as bit rate,traffic capacity,coverage,service availability and sustainability,etc.,and even the same indicators will have a more detailed division.All of these require the network to provide more detailed customization and scenario-based services,so business indicators need to be more finely divided.Aiming at the problem of access switching and resource scheduling in 6G fine-grained slicing,this thesis mainly studies the 6G fine-grained resource slicing and scheduling algorithm from the following three aspects:(1)Aiming at the network switching problem of fine-grained slicing of large bandwidth services,a 6G network elastic switching algorithm based on dichotomograph matching is proposed.The algorithm maximizes the total rate of users by completing the optimal correlation of "user-slice-base station",and models the access problem as a hierarchical dichotomous stable matching problem.The algorithm adopts two matching ideas of "slice-base station" integration and "first slice and then base station",and uses the Gale-Shapley matching algorithm to realize the adaptive switching of the network.At the same time,the algorithm considers the idea of elastic access,and users can choose any slice access that can meet business needs,and give preference to providing slices with high user rate.Simulation results show that the access success rates of the proposed integrated and two-stage matching algorithms are 15% and 10% higher respectively than that of the traditional method,and the total rate of users is also significantly improved.(2)Aiming at the problem of task offloading and slicing resource scheduling under multi-service,a joint algorithm of task unloading and resource allocation based on A3 C is proposed.Considering the fine-grained slicing scenario of multi-user and multi-MEC,task delay and energy consumption are taken as system costs to minimize system costs and complete task offloading and resource allocation.The slicing dimension is included in the task offloading decision,and the complexity of offloading decision and resource allocation increases,and it is difficult for traditional algorithms to meet the solution requirements.Therefore,using the A3 C algorithm,each MEC is trained locally as an agent through asynchronous parallelism,and the results are updated to the global network to accelerate the convergence speed.The simulation results show that this algorithm can reduce the system cost under the premise of meeting user needs.(3)A service-driven energy-saving and resilient network system is designed and implemented,which collects user,location and other information through the core controller to realize the control of base stations and users.When a user switches services,the system automatically selects the most suitable base station and slice for access,and hibernates the unused base station to reduce unnecessary resource consumption and achieve energy-saving effects.When a service is required,the system automatically wakes up the corresponding base station to achieve efficient use of network resources.The system design can be applied to various scenarios,which can effectively reduce network energy consumption and improve the utilization of network resources.The test results show that the system can realize the user’s adaptive access to the base station and slice,and the adaptive activation and sleep of the base station...