Electric Vehicle Cluster Assisted Multi-Tier Vehicular Edge Computing System: Cross-System Framework Design and Optimization

Cooperative vehicle infrastructure is critical for facilitating multi-tier vehicular edge computing (VEC) systems to collaboratively process latency-sensitive intelligent tasks. However, the dynamic topology and the scarcity of available computing resources on real traffic roads pose significant challenges. While previous works have explored using public transit and parked vehicles as cooperative vehicles, they often overlook the incentive mechanisms and the specific tasks of these vehicles, rendering these solutions impractical. Motivated by the achievements in smart electric vehicles (EVs), this paper presents an electric vehicle cluster (EVC) assisted multi-tier VEC system that leverages the capabilities of the vehicle-to-grid (V2G) technology to provide stable computing resources to traffic areas. The proposed EVC-assisted multi-tier VEC system employs a cross-system two-level optimization method that jointly considers offloading and resource allocation decisions for road task vehicles (RTVs) in the VEC system and charging power decisions for the EVs in the EVC. At the lower level, we minimize the task latency for RTVs by utilizing the computing resources of idle EVs. At the upper level, we build a multi-agent two-step game model and introduce a potential game-based strategy and a Nash equilibrium (NE) based EV-RSU mapping method to derive the optimal decision strategy for the EVC. Extensive simulation results demonstrate the efficient reduction in total-task latency of RTVs and lower cost for the EVC while meeting the essential requirements in the V2G.