A Comparative Methodology for Estimating the Capacity of a Freeway Section

The random characteristics of the traffic flow make it essential to have a random component and therefore add a stochastic meaning to the deterministic parameters. This paper aims to improve the conventional deterministic approach to the freeway capacity by estimating parameters of the various probability distribution functions that are likely to represent the probabilistic nature of freeway traffic capacity. Firstly, Maximum Likelihood Estimation method is applied to estimate the capacity distribution function. Then, confidence intervals for the capacity distribution function are calculated using Bayesian statistics techniques that can address the difficult problem of censored data. Finally, a comparative analysis has been conducted between the estimations of deterministic and probabilistic models to come up with a conclusion regarding spatial and temporal characteristics of freeway capacity. The analysis results indicate that including stochasticity in the model estimation results in better representation of observed data and thus improve understanding of real-life situations.