Research On Task Offloading Strategy Of Mobile Edge Computing In Heterogeneous Internet Of Things

With the explosive growth of mobile Internet and Internet of things(Io T)traffic demands,massive Io T devices have put forward higher-level requirements for network indicators such as channel bandwidth,latency,reliability,and user access.At the same time,the demands of computing-intensive tasks also pose challenges to the computing power of the device itself.On the other hand,the vigorous development of Io T makes the service object of the network no longer a single device or business scenario,and network heterogeneity has become an inevitable development trend.Therefore,in order to achieve differentiated control of different business types and flexible on-demand allocation of computing resources,it is necessary to deploy diverse network architectures to deal with complex and heterogeneous task scenarios.Aiming at the heterogeneous Io T scenarios based on Mobile Edge Computing(MEC),this paper focuses on two directions of terminal energy saving and delay optimization,and studies task offloading and resource allocation in computing offloading.The main work includes the following two aspects:(1)For the multi-user and single-server MEC task partial offloading scenario,study the weighted sum minimization of terminal energy consumption under resource and delay constraints.The original problem is decomposed into two sub-problems of individual unloading ratio and system resource allocation,and the Lagrange duality function is solved by KKT conditions,and an efficient and energy-saving algorithm for joint communication and computing resources is proposed.In the resource allocation part,a method based on projected sub-gradient is adopted to update the multipliers.At the same time,Dynamic Voltage and Frequency Scaling(DVFS)technology is adopted on user equipment to reduce unnecessary terminal energy consumption.A Software-Defined Networking(SDN)controller is used in the network to dynamically collect the computing power and channel bandwidth of each node in the network in real time.The simulation results show that the proposed algorithm can save 38.3% of the user’s equipment energy consumption compared with the traditional search algorithm.(2)For the multi-user and multi-cell MEC task offloading scenario,the problem of task delay weighted sum minimization under resource constraints is studied.A heterogeneous network architecture with edge-cloud collaboration is adopted,which makes full use of the low latency of edge computing and high computing power of cloud computing to solve the problem of latency optimization in computing offloading.Aiming at the 0-1 integer programming problem,a task offloading and resource allocation strategy based on Genetic Algorithm(GA)is proposed.In terms of task differentiation,a task pre-estimation model based on the Analytic Hierarchy Process(AHP)is proposed.The task data volume,the number of CPU cycles and the maximum task completion time are input into the model as parameters to obtain the task priority weight.The simulation results show that the proposed algorithm has good convergence,and compared with the traditional heuristic algorithm,it can effectively reduce the user’s average task processing time and improve the system resource utilization.