Reliability-oriented edge computation offloading strategy for Internet of Vehicles: based on improved hybrid fox optimization algorithm

Abstract In the Internet of Vehicles (IoV), edge computing supports computation-intensive tasks through roadside unit servers. However, guaranteeing reliable task execution remains challenging in dynamic environments with limited resources. We propose a reliability-oriented computation offloading strategy based on the Improved Hybrid Fox Optimization (IHFOX) algorithm. The algorithm integrates sine chaotic mapping and Lévy flight to balance exploration and exploitation. It further employs a standard normal distribution for position updates, thereby accelerating convergence and avoiding local optima. For model construction, we establish a single-replica reliability model from dual dimensions of transmission reliability and computational reliability, meeting system-level reliability constraints through a multi-replica parallel redundancy mechanism. Meanwhile, vehicle mobility, resource/energy constraints, and joint weight factors are integrated to form a comprehensive optimization model under reliability constraints. Experiments demonstrate that the IHFOX algorithm guarantees high reliability with over 95% task completion rate, while significantly reducing system cost, delay, and energy consumption, exhibiting superior practical value in reliability-sensitive IoV applications.