Research On Resource Allocation Algorithms For 5G Network Slices In Smart Grid

Driven by growing requirements of electricity and energy,as the essential infrastructure of modern society and the energy support of national development,the power grid has stepped into the era of smart grid from the traditional power grid.Various differentiated power business in smart grid have different requirements for network throughput,delay,connection number and reliability.As the key technology of 5G network,network slicing can tailor-made "exclusive network" for different power business to provide better service when facing the demand of network connection in a diversified vertical industry such as smart grid.Resource allocation of network slice is the most important factor to realize flexible network on demand.Efficient resource allocation algorithm can improve network resource utilization,guarantee user service quality and enhance network performance.The main contributions of this paper are summarized as follows:Firstly,the business requirements in smart grid are analyzed and summarized.The concept of business priority is introduced according to each business characteristic.Secondly,in view of the singleness of resource allocation and the neglect of fairness,the utility function is defined and a dynamic resource allocation strategy based on utility maximization algorithm for smart grid scenarios is proposed.Under the principle of fairness,the strategy realizes the joint optimization of user experience quality and system throughput.Thirdly,the feasibility and advantages of reinforcement learning in resource allocation are analyzed.A network slice resource allocation strategy based on reinforcement learning algorithm in smart grid is proposed,which achieves multi-objective optimization of spectrum resource utilization ratio and user service quality.The effectiveness of the proposed algorithm is verified by software simulation.Based on development status of 5G technology,research situation of smart grid and the future demand of power business terminals,reinforcement learning algorithm and utility maximization algorithm to realize the dynamic allocation of limited resources in smart grid are proposed in the paper.The algorithms meet the quality of user experience and business service requirements,conform to the evolution route of 5G technology and the development direction of smart grid system.The works in paper provide a reference scheme for the allocation of network slice resources in smart grid.