Efficient Resource Management in IoV through Task Offloading Optimization

With advancements in the Internet of Vehicles (IoV) and wireless technologies, intelligent vehicles are becoming increasingly prevalent, introducing applications that are computationally intensive, energy-consuming, and highly sensitive to latency. However, the limited computational capabilities of vehicles often result in excessive workloads that exceed local processing capacity, undermining real-time performance. VEC addresses this challenge by deploying computational resources at the network edge, particularly in Roadside Units (RSUs). In the field of task offloading for the IoV, recent studies have demonstrated the potential of Reinforcement Learning (RL) in addressing task offloading challenges, particularly in reducing latency and energy consumption. However, research on effectively utilizing computational resources within the IoV to further optimize latency and energy efficiency remains relatively limited, making it a promising area for future exploration. In this context, we propose a reinforcement learning-based optimization algorithm, PPOH, to enhance resource utilization and performance in the IoV environments. The algorithm optimizes the transmission power between vehicles and RSU, vehicle CPU frequencies, and RSU resource allocation ratios, balancing energy consumption and latency. Simulation results validate the convergence of PPOH, demonstrating its superior performance compared to benchmark algorithms across key metrics.