Energy-Aware and Mobility-Driven Computation Offloading in MEC

Heuristic algorithms are widely used in Mobile Edge Computing(MEC) to improve the performance of mobile devices. However, the time complexity of the heuristic algorithm is high, and it is complex to optimize under constraints. Therefore, we proposed Multi-User Energy Constraint Time Optimization Algorithm(MU-ECTOA) for workflow makespan optimization under energy constraints. MU-ECTOA includes three stages: cluster analysis, evaluation of performance, and workflow offloading. In the first stage, the workflow tasks are classified according to their characteristics; In the second stage, the subtask groups are obtained, and the evaluation results of each subtask group are obtained. In the third stage, the optimal subtask group is selected for offloading and then updated the ready time of edge nodes. Extensive experiments have been conducted, the ACO&GA, Max-Min, Particle Swarm Optimization(PSO), GA-DPSO, and SFLA are taken as the compared methods. The results of MU-ECTOA performs better in aspects in completion time, load balancing, and successful offloading rate compared with other methods. By comparing the results of algorithms, the makespans of the algorithms are close, but the algorithm complexity and load balancing of MU-ECTOA are much better; The time complexity of the MU-ECTOA algorithm is close to the Max-Min's, but MU-ECTOA performs better in makespan and algorithm reliability.