No abstract

As a novel computing technology closer to business ends, edge computing has become an effective solution for delay sensitive business of power Internet of Things (IoT) and promotes the application and development of the IoT technology in smart grids. However, the inherent characteristics of a single edge node with limited resources may fail to meet the delay requirements for access ubiquitous IoT businesses of massive access. Multiple edge nodes are needed to cooperate with each other to optimize workload allocation to provide lower delay services. To this end, this paper proposes a workload allocation mechanism, orienting edge computing-based power IoT, which minimizes service delay. The workload optimization allocation model is established, and the optimal workload allocation oriented on delay among multiple edge nodes is further realized on the basis of computing resource optimization within the single edge node. The balanced initialization, resource allocation, and task allocation (BRT) algorithm are proposed. Based on the balanced initialization of workload within edge nodes, the particle swarm algorithm modified by the pheromone strategy is used to solve the problem of the computing resources' allocation inside edge nodes. Finally, the task allocation among multiple edge servers is converted into a semi-definite programming problem. The simulation results show that the proposed BRT algorithm reduces the service delay by 9.1%, 16.9%, and 26.4%, and the service delay growth rate by 24.6%, 34.5%, and 38.7%, respectively, compared with the simulated annealing algorithm (SAA), LoAd Balancing (LAB), and Latency-awarE workloAd offloaDing (LEAD) algorithms. INDEX TERMS Edge computing, multiple business, power Internet of Things, service delay, workload allocation.

No abstract

No abstract

Abstract:The implementation of demand response (DR) could contribute to significant economic benefits meanwhile simultaneously enhancing the security of the concerned power system. A well-designed carbon emission trading mechanism provides an efficient way to achieve emission reduction targets. Given this background, a virtual power plant (VPP) including demand response resources, gas turbines, wind power and photovoltaics with participation in carbon emission trading is examined in this work, and an optimal dispatching model of the VPP presented. First, the carbon emission trading mechanism is briefly described, and the framework of optimal dispatching in the VPP discussed. Then, probabilistic models are utilized to address the uncertainties in the predicted generation outputs of wind power and photovoltaics. Demand side management (DSM) is next implemented by modeling flexible loads such as the chilled water thermal storage air conditioning systems (CSACSs) and electric vehicles (EVs). On this basis, a mixed integer linear programming (MILP) model for the optimal dispatching problem in the VPP is established, with an objective of maximizing the total profit of the VPP considering the costs of power generation and carbon emission trading as well as charging/discharging of EVs. Finally, the developed dispatching model is solved by the commercial CPLEX solver based on the YALMIP/MATLAB (version 8.4) toolbox, and sample examples are served for demonstrating the essential features of the proposed method.

As a mode of processing task request, edge computing paradigm can reduce task delay and effectively alleviate network congestion caused by the proliferation of Internet of things(IoT) devices compared with cloud computing. However, in the actual construction of the network, there are various edge autonomous subnets in the adjacent areas, which leads to the possibility of unbalance of server load among autonomous subnets during the peak period of task request. In this paper, a deep reinforcement learning algorithm is proposed to solve the complex computation offloading problem for the heterogeneous Edge Computing Server(ECS) collaborative computing. The problem is solved based on the real-time state of the network and the attributes of the task, which adopts Actor Critic and Policy Gradient's Deep Deterministic Policy Gradient(DDPG) to make optimized decisions of computation offloading. Considering multi-task, the heterogeneity of edge subnet and mobility of edge devices, the proposed algorithm can learn the network environment and generate the computation offloading decision to minimize the task delay.The simulation results show that the proposed DDPG-based algorithm is competitive compared with the Deep Q Network(DQN) algorithm and Asynchronous Advantage Actor-Critic(A3C) algorithm. Moreover, the optimal solutions are leveraged to analyze the influence of edge network parameters on task delay. INDEX TERMS Edge computing, computation offload, collaborative computing, reinforcement learning, DDPG.

No abstract

No abstract