Dynamic traffic light timing optimization model using bat algorithm

Abstract

Nowadays, traffic congestion is a crucial issue for many urban or even suburban areas. A potential but manageable cause of this effect is the inefficient control of traffic light timing cycle that is not proactively respond to the characteristics of the considering intersection and time-dependent condition of the traffic. In this paper, we propose a finite-interval model that finds effective green time for each of the four phases light timing in each cycle. The objective is to acquire satisficing solution, by means of Bat algorithm (BA) to minimize spoilage time at an intersection. The effectiveness of model was validated using a simulation experiment on a crowded intersection in Bangkok that has imbalanced numbers of accommodated lanes and formless urbanism. The traffic condition is simulated by the microscopic traffic simulator called `Simulation of Urban Mobility' (SUMO). The experimental result confirmed the effective improvement when compared with the traditional fix-time operational control.