Research On Dynamic Task Offloading And Resource Allocation Of Mobile Edge Computing System Based On Lyapunov Method

The advent of the 5G era and the popularization of the Internet of Things have opened the prelude to an era.The vision of interconnection of all things,information coming with one's heart,and everything within reach has become everyone's common expectation.In order to overcome the limitations of mobile devices in performing computationally intensive workloads,mobile edge computing(Mobile Edge Computing,MEC)can effectively solve the problem of insufficient energy and computing capabilities of Io T devices,and has become a key technology for next-generation networks.In recent years,the task offloading and resource allocation schemes in edge computing systems have received widespread attention from scholars.However,many of them only focus on static offloading schemes and do not consider the long-term average performance of the system.In order to optimize the long-term average performance,the randomness of task arrival,the dynamic change of the wireless channel and the mobility of the terminal,the research on dynamic offloading and resource allocation is full of challenges.This paper uses the Lyapunov stochastic optimization method to convert the long-term average problem into a single-slot deterministic problem.Without knowing the future system information,combined with the dynamic wireless channel,the arrival of random tasks,and the user's mobility,we have studied multiple problems.The service cost minimization and energy efficiency minimization issues under the server multi-user MEC system.First,the problem of unloading and resource allocation based on minimizing service cost in the multi-user and multi-server MEC system is studied.The service cost is defined as the weighted sum of power consumption and service migration cost.By jointly optimizing the CPU cycle frequency,transmit power and user correlation vector of mobile Io T devices(Mobile Io T Devices,MIDs),the long-term average and service cost of all MIDs are minimized.Aiming at this research goal,an online mobile sensing offloading and resource allocation algorithm based on Lyapunov optimization and Semi-Definte Program(SDP)is proposed.The simulation results show that the proposed scheme can balance the system service cost and delay performance,and is superior to other offloading benchmark methods in terms of system service cost.Then,the problem of offloading and resource allocation based on minimization of energy efficiency in the scenario of multi-user and multi-server MEC is studied.Taking user mobility,dynamic wireless channel state,and random task arrival into account,the energy efficiency performance index is defined.The optimization problem is established to ensure the stability of the task queue,as well as the problem of minimizing energy efficiency under the constraints of peak transmit power,maximum CPU cycle frequency and maximum number of user associations.The optimization problem of this design is that the non-convex problem is difficult to solve.To this end,this paper first uses Dinkelbach's algorithm to convert the nonlinear fractional programming problem into a linear fractional programming problem,and then proposes an online mobile sensing energy-saving and efficient offloading algorithm based on the Lyapunov stochastic optimization and sub-model method to solve the offloading decision and resource allocation problems.Theoretical analysis proves that the proposed algorithm can achieve a balance between system energy efficiency and service delay [(46)(1/ V),(46)(V)].Simulation results verify the theoretical analysis and prove that compared with several benchmarks,our proposed algorithm can provide better energy efficiency-delay performance.