Dynamic Offloading Method for Mobile Edge Computing of Internet of Vehicles Based on Multi-Vehicle Users and Multi-MEC Servers

With the continuous development of intelligent transportation system technology, vehicle users have higher and requirements for low latency high service quality task computing. The computing offloading technology mobile edge (MEC) has received extensive attention in Internet Vehicles (IoV) architecture. However, due to limited resources MEC server, it cannot meet requests from multiple simultaneously. For this reason, making correct fast decisions provide with a latency, energy consumption, cost is still considerable challenge. Regarding issue above, IoV environment where race, paper designs three-layer overhead model based on Edge-Cloud collaboration servers. To solve problem minimizing total performing tasks, an collaborative, dynamic computation method (ECDDPG) deep deterministic policy gradient designed. This deployed at layer make tasks generated by users. simulation results show that performance better than Deep Q-network (DQN) Actor-Critic regarding reward value convergence. In face change wireless channel bandwidth number users, compared basic strategy, proposed reducing computational cost, delay, consumption. At same time, complexity execution significantly reduced.