عنوان مقاله [English]
Accurate estimation of delay is difficult because of the randomness of traffic flow and large number of factors affecting intersection capacity. Existing delay models provide only approximated point of estimates of signalized intersections. Identification and evaluation of probability distribution of delay can be beneficial in understanding the nature of its variability at signalized intersections. In this study travel time delay data at three per-timed signalized intersections in Tehran, Iran were evaluated. The main purpose of this research is to present a delay distribution prediction model based on the data collected at some signalized intersections in Tehran. For this purpose the following steps have been taken:
1. Literature reviews on queuing theory and existing delay models, well as well as theoretical support for the study.2. Sampling and data collection regarding delay of vehicles in some fixed time signalized intersections in Tehran.3. Determination of delay distribution functions by existing vehicles delays values obtained from the data. For this purpose the frequency of delays (data) in various intervals was calculated and then, based on the obtained histogram of the data, a few distribution functions (i.e. Weibull, Gamma, Beta, Normal, Erlong, Triangular, Exponential and Uniform) were fitted to this data and consequently the best delay distribution function was obtained.4. Presentation of a model for predicting the delay time in some signalized intersections in Tehran.
Hence, delay probability distribution functions were developed for each signalized intersection using Arena simulation software. Results of simulation show that Weibull distribution function provided better fits for delay data at all three intersections. The two parameters of this distribution were estimated by "average delay" and "g/c" ratio.