Conventional Optimization Technique Based Methodology to Estimate Gap Acceptance Functions

Abstract

The definition of a critical gap in literature supports that it not only varies from driver to driver but also for (over) rejected gaps by a driver. In this light, multinomial probit (MNP) models were proposed in the past as gap acceptance functions. There are few methodologies available in the literature to estimate these functions. These methodologies are either computationally expensive or involve sophisticated mathematics. This is possibly one of the reasons due to which MNP models are not explored extensively as a tool to model critical gaps. Further, estimation of such gap acceptance functions always leads to maximization of likelihood functions. This paper proposes a methodology based on a conventional optimization technique for the first time to estimate the parameters of a class of gap acceptance functions. It is also shown that the estimation results match with those from another method, namely CHOMP. The proposed method turns out to be a simple yet effective one.