A task offloading strategy based on sequential waiting model in MEC

In the field of mobile edge computing (MEC) research, many studies under common research scenarios focus on the optimization of energy consumption and processing delay. Meanwhile, some researchers only discussed the task offloading of terminals to a single server, without considering the collaborative execution of task offloading by multiple servers. Motivated by their observations, this paper conducted research in the scenario of offloading tasks from a single terminal to multiple edge servers. We formulated the multi-objective optimization problem of latency and offloading reliability, and proposed an intensive task offloading strategy based on the Sequential Waiting Model(SWM). The scheme allocates tasks according to the wireless channel environment and the computing power of the server. Besides, this approach minimized the combined cost consisting of latency and offloading failure probability. Since this kind of optimization problem has been proved to be an NP-hard problem. And we designed a Niching Preservation Genetic Algorithm (NPGA), which is based on Niching Genetic Algorithm (NGA). Besides, to obtain better system performance, we simulated the influence of different computing factors on the proposed algorithm NPGA and analyzed the convergence of the NPGA. Finally, the simulation results illustrated the proposed algorithm NPGA can effectively reduce the total latency of task completion and the probability of task offloading failure. Compared with the most advanced algorithms in the previous literature, the cost function value of the combined latency and offloading failure probability reduces 2%-16%. At the same time, we conducted simulation experiments using real base station locations, and also verified that the NPGA algorithm can achieve lower cost values under multiple edge server counts.