Computation Offloading and Resource Allocation in Failure-Aware Vehicular Edge Computing

The advent of Intelligent Cyber-Physical Transportation Systems (ICTS) has not only accelerated the reformation and evolvement of smart transportation, but also ushered in a new era of vehicular applications. These applications typically impose stringent latency requirements and demand substantial computing resources. Vehicular edge computing (VEC) has emerged as an efficient solution to address these challenges, leveraging its inherent ability to provide ultra-low latency services. Existing studies primarily concentrate on either optimizing resource allocation or minimizing response latency, while ignoring the fact that the task execution in VEC is more susceptible to failures compared to cloud computing environments. Accordingly, we design a cost-efficient and failure-resistant task offloading strategy for VEC systems with the goal of minimizing the average response latency for all tasks. Specifically, our problem is modeled as a nonlinear multi-constraint continuous optimization problem, with tightly coupled optimization variables in the objective function and constraints. To tackle this issue, we initially decompose the optimization problem into per-slot optimization subproblems. Subsequently, we employ an effective algorithm with low time complexity to solve these subproblems in a slot-by-slot manner. We comprehensively evaluate the performance of our approach through extensive simulations, demonstrating that our method outperforms the baseline approaches in various aspects.