A Comparative Assessment of Stochastic Capacity Estimation Methods

Abstract

The stochastic nature of highway capacity has gained increasing attention in recent times. For the empirical estimation of capacity distribution functions based on measured traffic data, two methodologies have received considerable application: The direct estimation of breakdown probabilities for groups of traffic volumes on the one hand and the estimation of capacity distribution functions based on statistical models for censored data on the other hand. The objective of the paper is to compare these methods in terms of estimation accuracy, applicability, and consistency of the results. The theoretical differences of both methods as well as the consequences for application are discussed and analyzed based on empirical traffic data as well as data from macroscopic simulation. The analysis yields that the capacity estimation based on models for censored data performs better than the direct breakdown probability estimation technique, particularly concerning the consistency of the estimated capacity distribution functions.