Elvin J Moore, Wisut Kichainukon, and Utomporn Phalavonk

pp. 489 - 499

Abstract

A road network can be modeled as a graph with a set of nodes representing intersections and a set of weighted edges
representing road segments between intersections. In this paper, a traffic flow problem is studied, where edge weights
represent road capacities (maximum vehicles per hour) that are functions of the traffic speed (km/hr) and traffic density
(vehicles per kilometer). To estimate road capacities for a given speed, empirical data on safe vehicle separations for a given
speed are used. A modified version of the Ford-Fulkerson algorithm is developed to solve maximum flow problems with speeddependent capacities, with both one-way and two-way flows allowed on edges and with multiple source and target nodes.
The modified algorithm is used to estimate maximum traffic flow through a selected network of roads in Bangkok. It was found
that the maximum safe traffic flow occurs at a speed of 30 km/hr.