Adaptive Task Offloading and Resource Management for Vehicular Edge Computing

Within the field of dynamic vehicular networks, also known as vehicular edge computing (VEC), task offloading and resource allocation face significant challenges due to constantly changing network conditions and limited resources. To address these issues, we propose a novel method for task offloading and resource allocation that considers the unpredictable starting points of vehicle offloading and the immediate impact of vehicle movement on wireless communication links. We utilizes dynamic channel modeling to track the fluctuations in link state, incorporating a task classification and migration strategy. This approach allows us to adaptively modify both single and cross-region offloading. Furthermore, to minimize overall latency and energy consumption, we optimally coordinate communication and computing resources by dynamically assigning vehicle transmission power, local computing resources, and the computing resources of roadside units. Simulation results indicate that the proposed approach effectively minimizes both latency and energy consumption. It notably outperforms benchmark schemes, showing a significant reduction in the Z value, achieving lower task delays and reduced energy consumption compared to existing strategies. These findings highlight the substantial advantages of the proposed approach, particularly in dynamic vehicular environments.