Energy-Efficient Multi-Vehicle Edge Networks: A Joint Optimization Algorithm for Task Splitting Offloading and Resource Allocation

Abstract

In Internet of Vehicles (IoV), considering the different task volumes, the changing trajectories and the limited resources of the RSU server, it is a major challenge to effectively coordinate task offloading between multiple vehicles and reduce energy consumption. Therefore, to solve the problem of numerous intensive tasks in a multi-vehicle RSU system, we propose an vehicular edge computing (VEC) optimization scheme for joint the computation offloading, unequal task splitting and resource allocation. The goal is to minimize energy consumption with guaranteed task completion latency and limited RSU computing resources. The range of movement of each task vehicle and the optimal vehicle computational resource allocation are considered as well. As the joint optimization scheme is mixed-integer nonlinear optimization problem (MINLP), it is decomposed into two sub-problems. Firstly, given the initial CPU frequency allocation for each vehicle task, the collaborative offloading strategy is determined using game theory. Secondly, taking into account the constraints of computing resources, a sequential quadratic programming (SQP) algorithm is used to provide optimal strategies for unequal task splitting and CPU frequency allocation. Simulation results show that the proposed algorithm can reduce the system energy consumption by about 30% compared with the Game Theory-based scheme under the same latency constraint.