Load-Aware Multi-Task Offloading Strategy Based on Optimal Stopping Theory in Mobile Edge Computing

With the wide application of Mobile Edge Computing (MEC) in the Internet of Vehicles and other fields, mobile nodes are no longer limited to smart phones, but also include new mobile devices such as smart vehicles. Faced with the access of a large number of devices, there may be situations where some edge servers are overloaded while others are relatively idle. To address this issue, a load aware multi-task offloading strategy based on optimal stopping theory is proposed, which minimizes the average load of the selected edge servers through sequential decision-making. Firstly, consider the cost of delaying decision-making and introduce a benefit function to constrain the number of observations. Secondly, the optimal stopping time is determined by solving the secretary problem with observation constraints. Finally, extending to the scenario of multi-task offloading, design a phased decision-making multi-task offloading algorithm. The simulation experiment results show that the proposed offloading strategy can significantly reduce the average load of the selected edge servers for task offloading, and can more effectively optimize the workload of edge servers.