Differential pricing based task offloading scheme for low latency application in MEC system

Abstract

This study deals with pricing scheme based on the user's usage of server computational resources in situations where users handling low-latency applications partially offload to meet task completion time requirements and the server gains revenue from it in a mobile edge computing environment. It also addresses scheme that determines the optimal usage of server computational resources and the amount of data to offload from the user's perspective, taking into account the charges based on the proposed pricing scheme and the user's task completion time. Recently, studies have been conducted on pricing scheme considering only the amount of offloaded data to the server. However, according to this approach, the user's usage of server computational resources does not affect the payment, so all users try to use lots of server computational resources. Since the server's computational resources are finite, to support offloading for many users, it is necessary to operate efficiently by receiving more money from users who use more computational resources. To solve this problem, this study proposes a differential pricing scheme in which unit price per second are determined based on the user's usage of server computational resources, and derives the user's optimal usage of server resources and the amount of data to offload through the convex optimization problem considering the payment based on proposed pricing scheme and the task completion time. Through numerical results based on simulations, it is shown that the proposed pricing scheme and the user's offloading decision scheme can solve the problem mentioned above and increase the efficiency of mobile edge computing servers.