An Improved ARO Model for Task Offloading in Vehicular Cloud Computing in VANET

Abstract

Abstract Vehicle-to-vehicle (V2V) communication enables a network of automobiles to perform collaborative computing, giving rise to the concept of a "vehicular cloud" (VC). However, without the need for edge nodes or cloud servers, vehicles can carry out applications needing the massive amount of processing cooperatively on their own by creating a Vehicular Ad-Hoc Network (VANET). Managing the recurrent topology alteration caused by vehicle mobility is a significant challenge for VANET cooperative computing. In this research, we present a V2V-based cooperative computing approach. The suggested method takes into account the distance between vehicles while choosing which ones to collaborate with, and it waits task offloading until the last possible moment to ensure a stable and energy-efficient cooperative computing environment. Despite its competitive performance when compared to other MH algorithms, the artificial rabbits optimisation (ARO) algorithm still suffers from poor accuracy and the issue of becoming trapped in solutions. By antagonism methods, this research creates selective opposition version of the artificial rabbit procedure (LARO), which eliminates the negative consequences of these shortcomings. To begin, during the random concealment phase, a Lévy flight strategy is implemented to increase population diversity and dynamics. The algorithm's convergence accuracy is enhanced by the richness of its various population samples. The tracking efficiency is improved, and ARO is kept from getting stuck in its existing local solutions by adopting the selective opposition technique. In comparison to traditional static scheduling techniques, the suggested strategy improves upon both energy efficiency and network reliability.