Simulation is very useful for generating traffic data. However, for validating a traffic model, field data are essential. Because of the random variation of the traffic arriving at signalized intersections, a large number of observations are necessary at a particular intersection. To obviate the collection of a large amount of data, the Bootstrap technique was applied to a limited amount of field data that were collected at three fixed-time signalized intersections. In addition, the data were supplemented by simulation to cover a wide range of cycle lengths, types of flow, degrees of saturation, and ratios of the variation in the mean of the number of arrivals per cycle. A recently developed traffic model based on a Markov process and a geometric probability dis tr ibution, the Modified Geometric Model (M Geom Model), was used to estimate the measures of effectiveness commonly used for optimization purposes, namely, delay and stops. Satisfactory results were obtained, indicating that a limited amount of field data is required for validating a traffic model.

Traffic arrival at a signalized intersection is a stochastic process. If a traffic count is taken on a specific day for a particular period at an intersection and the delay and stops are calculated, when the count is repeated for the same time period the next day, it is possible that the values obtained for delay and stops may differ appreciably. Consider Set A of generated arrivals obtained over 15 cycles with the following conditions: cycle length (c) is 50 sec; degree of saturation (x) is 1.05; flow (Q) is 500 vehicles per hour; and ratio of variance in the mean of arrivals per cycle (I) is 0.5.