Task Offloading for Mixed Cloud/Fog Computing in Vehicular Network Using Genetic Particle Swarm Optimization

Intelligent connected vehicle (ICV) is considered to be a key technology to solve traffic safety and congestion problems. With the development of ICV technology, vehicles need to perform complex image processing and control decision with deep learning operations. These vehicular applications require a lot of computing power while computation capability of vehicle is limited. The joint application of cloud computing and fog computing in the vehicular network can provide vehicles with sufficient computing resources through reasonable task offloading scheme, and reduce high communication delay. The paper formulates a problem to minimize the total service latency of all tasks in the system, considering the effect of vehicle movement, communication bandwidth limitations, and task queuing time. A novel hybrid metaheuristic algorithm named improved genetic binary particle swarm optimization (IGBPSO) is proposed to solve this problem, which combines the operations of particle swarm optimization (PSO) and genetic algorithm (GA). Experimental results demonstrate that the algorithm has both fast convergence and strong global search capabilities, which can effectively reduce the total service latency for all tasks and improve the utilization of system resources.