Urban Traffic Flow Optimization using Intelligent Techniques

Abstract

This work describes optimization of traffic flow in an urban area using different intelligent techniques such as Evolutionary algorithm (EA), Simulated Annealing (SA) and a proposed novel Hybrid Evolutionary algorithm and Simulated Annealing (EASA) Technique. The proposed work focuses on communication between vehicles and road side unit (RSU) referred as Vehicle to Infrastructure (V2I) architecture. Vehicles store their locations and destinations in a central server. The central server is assumed to compute the best path for each of these vehicles using different intelligent techniques and vehicles receive information regarding the route that consumes least time. The fitness function minimizes the total traffic flow in an urban area by estimating the best possible routes for each of the vehicles and optimizing the total travel time. The Petri net model is used to validate the proposed work. In this work, a novel hybrid Evolutionary algorithm and Simulated Annealing technique referred as hybrid EA-SA is proposed to optimize the traffic flow in an urban area. The proposed hybrid EA-SA is validated with a traffic queue near a traffic junction. The traffic junction makes the vehicles wait for a certain fixed time interval or for a variable time interval and hence, creates a queue. The first part of the work focuses on optimization of traffic flow without any queue or waiting time near a junction. Furthermore, performance of the proposed approach is analyzed in presence of a traffic light controller near a junction with fixed and variable green time interval. The vehicles wait near a junction for a fixed time interval referred as fixed green time interval irrespective of the size of queue at a junction. Vehicles waiting time is also varied near a traffic junction referred as variable green time interval.