Sustainable Internet of Vehicles System: A Task Offloading Strategy Based on Improved Genetic Algorithm

Abstract

“Smart transportation” promotes urban sustainable development. The Internet of Vehicles (IoV) refers to a network with huge interaction, which comprises location, speed, route information, and other information about vehicles. To address the problems that the existing task scheduling models and strategies are mostly single and the reasonable allocation of tasks is not considered in these strategies, leading to the low completion rate of unloading, a task offloading with improved genetic algorithm (GA) is proposed. At first, with division in communication and calculation models, a system utility function maximization model is objectively conducted. The problem is solved by improved GA to obtain the scheme of optimal task offloading. As GA, in the traditional sense, inclines to a local optimum, the model herein introduces a Halton sequence for uniform initial population distribution. Additionally, the authors also adapt improved GA for the problem model and global optimal solution guarantee, thus improving the rate of task completion. Finally, the proposed method is proven through empirical study in view of scenario building. The experimental demonstration of the proposed strategy based on the built scenario shows that the task calculation completion rate is not less than 75%, and when the vehicle terminal is 70, the high-priority task completion rate also reaches 90%, which can realize reasonable allocation of computing resources and ensure the successful unloading of tasks.