The effect of random headways on throughput at an oversaturated signalized intersection is studied. The model assumes an infinite queue of cars at the beginning of the go-phase. The entry time of the first car into the intersection, after the go-phase is initiated, is assumed to be a positive random variable x, and all successive entry- time headways (minus a minimum time) are assumed to be the random variable x. Furthermore, all the random variables are assumed to be mutually independent. Analytic expressions are derived for the mean and the variance of the number of vehicles that pass through the intersection during a go- phase. Some calculations and curbs are given for the specific case of an erlangian distribution function of the random variable x. It is shown that a deterministic (mean value, model is quite accurate for mean value calculations, but that variability in headways does cause significant variance in throughput. /author/.
Abstract