Research On Scheduling Strategies Of User Energy Businesses Based On Cloud Platforms

In recent years,with the gradual maturity of clouds and edge clouds techniques,the cloud platform is able to realize more efficient resource management and to provide more effective guarantees of business latency and reliability,which makes it possible to build a more efficient platform for informationized control and management of power energy,especially in the aspects of distributed renewable energy generation(DREG),distributed power storage and electric vehicles(EVs).Additionally,with the increasing number and variety of intelligent terminals in the industry of power energy,the architecture of energy businesses is growing large and complex,which produces new requirements for efficient platform-based control.It is urgent and necessary to put forward strategies of energy business scheduling that are able to take advantages of computing and storage resources of clouds and edge clouds and to meet user requirements for low business latency,high energy utility,etc.In the existing power energy grids,there are many types of energy terminals,the network structure is complex,and the network management and control are difficult,so that it is required to design the system architecture of the user energy business platform at first.On this basis,it is supposed to further deal with the challenging issues,including the collaborative scheduling of large quantities of businesses,the low speed of business responses,the low resource efficiency and enormous energy consumption of the platform.To address these issues,this paper studies the strategy of user energy business scheduling in the edge cloud and the whole cloud platform,respectively.Additionally,as for the case of EV businesses,this paper studies a platform-based charging scheduling strategy for multiple EVs in order to deal with the challenges in joint optimization of charging time,charging location and driving routes of EVs.The main work and innovation of this paper can be summarized as following:1.Architecture design of the edge-cloud collaborative cloud platform of user energy businessesThe existing cloud platform of energy businesses is based on cloud data centers,which is of poor scalability,large transmission latency to users,high risks of single point failure,etc.To address these issues,this paper designs an edge-cloud collaborative architecture for cloud platforms of user energy businesses,where clouds and edge clouds are hierarchically deployed aiming at the flexible extension of cloud platforms,and an uniform architecture of cloud techniques is applied to the construction of edge-cloud collaborative schemes in the aspects of resource,data,business management,etc.Based on the proposed architecture of cloud platforms,a new cloud-edge-terminal collaborative framework of user energy management is put forward to realize the coordinated scheduling of renewable power generation,power storage and smart loads on user-sides.It innovatively uses a two-timescale framework to enable us to offload the fine-grained strategies of energy control,containing detailed informations like device types and scheduling time,from clouds to terminals on usersides,which is able to avoid the transmissions of crucial control information in networks as far as possible,and then to reduce the cyber-physical security issues of users caused by the loss,the delay of information,etc.It is shown by simulation results that the proposed edge-cloud collaborative architecture of cloud platforms can reduce the business latency by 0.3-0.5 seconds as well as provide more effective reliability guarantees compared with the existing architecture of cloud platforms,benefiting from the application of multi-node collaborative computation and the advantages that edge clouds are closer to users,which is helpful for real-time businesses.Additionally,the proposed cloudedge-terminal collaborative framework of energy management can reduce energy costs of users by 10%-30%under different configures of electricity price and their battery capacities.2.Strategies of charging scheduling for multiple EVs based on the cloud platformIn the case of EV businesses,the anarchies of charging time,charging location and driving routes of large quantities of EVs can lead to both the long charging time(or called charging latency)of users and the low resource efficiency of charging infrastructure.To address this issue,this paper proposes a platform-based strategy of charging scheduling for multiple EVs,which can recommend the latency-minimized charging stations(CSs)and the optimized routes to EV users according to their individual trip plans,and can also carry out efficient charging scheduling of multiple users according to the global view information for the purpose of improving the resource efficiency of charging infrastructure.At first,the problem of charging scheduling for multiple EVs is formulated by a graphical game model aiming to minimize the charging latency of users.Then,in the solution process,it is the first time to introduce the theory of correlated equilibrium(CE).CE is able to achieve a smaller sum of charging latency of EV population than that obtained by Nash equilibrium(NE),because it can obtain a more practical joint probability distribution of users’ mixed strategies.Additionally,CE can significantly reduce the problem complexity in the sense that it can be solved by linear programming techniques.According to simulation results,our proposed strategy of charging scheduling for multiple EVs based on the graphical game has apparent advantages of reducing charging latency of users and improving the resource efficiency of charging infrastructure compared with the existing methods,especially in cities with dense distributions of both EVs and CSs.3.Strategies of latency-minimized and privacy-aware scheduling of user energy businesses in edge cloudsThe large quantity,distributed deployment and huge differences in resource of edge-cloud entities result in significantly increased difficulties in collaborative scheduling of user businesses,which hinders the improvements of business response speed and resource efficiency of edge clouds.Additionally,the inadequacy in data security mechanisms of edge clouds is very likely to bring about data leakages and then to cause damages to user privacy.To address these issues,this paper proposes a business scheduling strategy in edge clouds based on the theory of completely potential game,which can realize the collaborative scheduling of multiple users according to the potential function of all users for the purpose of reducing the total business latency of user population and improving the resource efficiency of edge clouds,and can achieve privacy-aware task scheduling by restricting users from dispatching tasks to entities with high PLRs.Specially,to evaluate the PLRs resulting from the data leakages in different edge entities,the theory of sequential probability ratio testing(SPRT)is introduced to analyze the probability that users’ private features are correctly learned by adversaries,based on which the mapping function of the sample size and PLR can be built.Simulation results show that our proposed strategy of business scheduling in edge clouds can reduce the PLRs experienced by users compared with the existing methods like Gale-Shapley,and can further reduce the business latency of users and improve the resource efficiency of edge clouds by effectively balancing the workload among different edge entities.4.Strategies of latency and power balanced business scheduling in cloud platformsThe business latency and power cost of cloud platforms are usually opposed to each other.However,the existing strategies of business scheduling in cloud platforms lack flexibility in the trade-off and optimization of the two key indexes,and also have drawbacks of unstable business latency,insufficient considerations of clean energy,etc.To address these issues,this paper proposes a power cost and business latency balanced strategy of business scheduling in cloud platforms on the basis of a two-timescale framework.At first,in large time scales,a new utility function jointly optimizing the strategies of power purchasing,power storage and capacity provisioning of cloud data centers(CDCs)is designed,which can realize a weighted balance between business latency and power cost according to the index preference.Then,in small time scales,the dynamical scheduling of businesses from edge gateways to CDCs is realized on the basis of a framework of Lyapunov optimization.As a consequence,by respectively optimizing the capacity provisioning of CDCs and the business scheduling from edge gateways to CDCs in two different time scales,measured by hours and minutes,the frequent state switches of servers can be avoided efficiently,so that more stable business latency is available.Additionally,a pollution indicator function(PIF)is introduced to measure the pollution cost of power generated by different categories of energy in order to establish punishment mechanisms to encourage CDCs in the uses of clean power.Simulation results reveal that the proposed two-timescale strategy of business scheduling is able to realize the flexible trade-off between power cost and business latency of CDCs,and can achieve green uses of power by improving the usage proportion of clean power to 40%-60%under the incentive of PIF.