Optimal Task Offloading Strategy in Vehicular Edge Computing Based on Game Theory

In vehicular edge computing, when there are many vehicles requesting offloading services at the same time, relying only on the resources of edge servers often cannot meet the needs of delay-sensitive tasks. Most existing task offloading studies tend to only consider pure offloading strategies for vehicles, which may not be the optimal strategy for some splittable tasks. In this paper, we jointly optimize the vehicle hybrid offloading strategy and the server resource pricing strategy. For a requesting task, it can be executed locally, be offloaded to the edge server, and be offloaded to the cloud center at the same time. We model the interaction between vehicles, the edge server and the cloud center as a game model. Based on the analysis of backward induction, we prove that the game has a unique Nash equilibrium. Meanwhile, an algorithm that can converge to the equilibrium point in polynomial time is proposed. Numerical experimental results show that the proposed algorithm has better performance in terms of delay and cost than existing algorithms.