Research On Multi-objective Optimized Task Scheduling Algorithm In Mobile Edge Environment

With the rapid development of Mobile Edge Computing(MEC),the task scheduling problem in the MEC environment has raised attention.However,task scheduling is still immature in the research of MEC.On the one hand,unreasonable task offloading will cause the waste of time and energy,but existing scheduling algorithms do not consider the impact of network environment changes on task offloading,which is caused by the movement of user.On the other hand,in addition to the higher requirements for optimizing task delay and lower energy consumption of mobile devices,the task reliability is also highly required.However,existing scheduling algorithms lack consideration of multi-object joint optimization problems when reliability constraints are taken into account.This article conducts in-depth research on the two above problems,and the main contributions of this paper are as follows:1.This paper proposes a task scheduling algorithm on edge-side optimization in the MEC environment(D2C-OeNB),to solve the problem of unreasonable task offloading in the device to center(D2C)scheduling algorithm.The algorithm obtains an offloadable evolved NodeB(eNB)set with hidden Markov model,then combines speed and offset to select the optimal eNB to improve the success rate of offloading.Besides,during the task scheduling process of the algorithm,the strategy of dynamically updating the critical path is used to adjust the scheduling sequence,thereby further realizing the optimization of MD energy consumption and task delay.2.This paper proposes a task scheduling algorithm based on reliability constraints in the MEC environment(D2D-RC),to solve the problem that device to device(D2D)scheduling algorithms are lack of consideration of multi-object joint optimization problems with reliability constraints.The algorithm converts task load queue,MD energy consumption,and reliability constraints into a computing model,then uses Lyapunov optimization method to reconstruct the computing model as a problem with queue stability constraints.Finally,the branch and bound method is used to reduce the computational complexity of the online scheduling algorithms,so as to quickly find the global optimal scheduling solution.3.This paper simulates task scheduling under the MEC environment through EdgeCloudSim,and implements D2C-OeNB algorithm and D2D-RC algorithm.The experimental results demonstrate that the D2C-OeNB algorithm reduces energy consumption by 5.6%and 9.2%compared to PSwH and HEFT,and reduces time consumption by 12.6%and 13.1%compared to PSwH and HEFT,respectively.Compared with the DO-Greedy algorithm,the D2D-RC algorithm reduces queue length and energy consumption by 14.2%and 19.1%,respectively.