Performance Optimization by Using Artificial Neural Network Algorithms in VANETs

Abstract

The IEEE 802.11 standard provides specifications of medium access control (MAC) layer for VANET. The performance can be optimized by optimizing contention window size. In this paper, to optimize the performance in VANETs three different artificial neural network (ANN) algorithms are used to find the optimum contention window (CW) size which are Particle Swarm Optimization (PSO), Differential Evolution Algorithm (DEA) and Artificial Bee Colony Algorithm (ABCA). Performance comparison among PSO, DEA, ABCA and traditional MAC based on IEEE 802.11 is provided. The simulation results show that the ANN algorithms improve the communication reliability by increasing the throughput and decreasing the packet dropping rate (PDR).