Dependency-Aware Task Offloading and Resource Pricing in Parked Vehicle-Assisted Edge Computing: A Stackelberg-Contract Approach

Parked Vehicle-assisted Edge Computing (PVEC) significantly reduces task latency and energy consumption of user vehicles (UVs) by offloading their computation tasks to parked vehicles (PVs) with idle resources. However, uneven resource distribution and strong task dependencies in PVEC networks may lead to failed task offloading and unfair resource pricing. To address this issue, this paper proposes a Stackelberg game and Contract theory-based Dependency-aware task Offloading and resource Pricing (SCDOP) framework. Firstly, we construct a novel PVEC architecture, consisting of UVs, PVs, SDN controller, and roadside units (RSUs), where UVs partially offload their dependency-aware tasks to RSUs managed by SDN controller, and RSUs with inadequate resources will forward the tasks to nearby PVs. Subsequently, we depict the interaction between RSUs and UVs/PVs as a Stackelberg game/contract process, and prove economic properties of the SCDOP framework to ensure its effectiveness and fairness. Next, we formulate a social welfare maximization problem by jointly optimizing task offloading decision, resource allocation and pricing strategies, and propose a Particle Swarm Optimization Plus Genetic (PSOPG) algorithm to find approximate optimal solutions. Finally, simulation results demonstrate that the proposed PSOPG outperforms other benchmark schemes in terms of task completion rates and social welfare across various scenarios.