Satisfaction Optimization in Failure-Aware Vehicular Edge Computing

Abstract

Vehicular edge computing (VEC) has gained worldwide attention in both academia and industry. Current works on VEC mainly focus on task offloading and resource allocation to improve the performance of VEC systems, but seldom consider the satisfaction level of vehicles. Whereas, the satisfaction level of vehicles has been playing an important role in stimulating vehicles to pursue better quality of experience by task offloading and service outsourcing operations. In the meanwhile, there is an inescapable fact, i.e., the task execution in VEC may fail due to various reasons, and thus it is important to incorporate the failure-resisted task offloading into the failure-prone VEC system. In this paper, we aim to maximize the satisfaction of all the vehicles, while considering the potential failures in VEC. Specifically, we model satisfaction optimization as a multiple knapsack problem and further put forward a greedy heuristic approach to solve this problem in polynomial time. Extensive simulation is carried out to validate the efficiency of our approach in terms of the optimal values and the running time. The simulation results have shown that our approach can achieve a better result compared to other benchmarks.