Particle Swarm Optimization based Task Offloading in Vehicular Edge Computing
Niharika Keshari, Tejas Subhashchandra Gupta, Dinesh Singh
Abstract
Recent development in the field of Internet-of-Things (IoT) enhancing various safety and non-safety services of Vehicular Ad-Hoc networks (VANET). These applications execute on cloud or Mobile Edge server (MES) due to resource restriction of IoT devices. MES is limited in-network because of higher installation costs. Hence, some applications are unable to get service within the deadline. To resolve this, a new evaluation has been introduced named Vehicular Edge Computing (VEC) which utilizes idle vehicles as an edge server to fulfill upcoming demands of IoT. The challenge in VEC is to offload the task to the appropriate vehicle according to available resources, position and speed in order to improve resource utilization and makespan. Hence in this paper, we have proposed an offloading mechanism that offloads tasks using Particle Swarm optimization (PSO). Here the PSO assigns the best fit task to the vehicle according to position and speed of the vehicle for completing the computation within time. Simulation using OMNET++, Veins, and Sumo, shows that the proposed PSO-based offloading improves makespan, resource utilization and offloading ratio around 28.1%, 17.43% and 47.75% compared to Branch and Bound offloading technique.