Simulation of Garland, Texas, vehicular traffic using current and computed optimal traffic settings

摘要

This paper presents results of a study utilizing computer simulation of vehicular traffic in the downtown area of Garland, Texas. A general discrete digital simulation model, the Vehicle Traffic Simulator (VETRAS), developed by IBM Corporation, was used for the simulation. Using data supplied by the City of Garland, traffic patterns for three peak periods of daily operation---A.M., Noon, P.M.---were simulated. Two simulations were run for each period. In the first, intersections were controlled with signal settings currently in use in Garland. In the second, intersections were controlled by signal settings derived via a pattern optimization algorithm. A minimum interference technique was used to compute coordinated signal settings and offsets to maximize arterial and network performance.

The results show that arterial and network performance improves in each of the peak periods using the computed signal settings. Further, there is a direct relationship between volume and relative improvement. Estimated cost benefits for these improvements are also presented.