Modeling effects of travel time reliability on mode choice using cumulative prospect theory

• Published in: Transportation research. Part C, Emerging technologies Emerging technologies, 2019-11, Vol.108, p.245-254
• Main Authors: Ghader, Sepehr, Darzi, Aref, Zhang, Lei
• Format: Article
• Language: English
• Subjects: Cumulative prospect theory; Mode choice; Observational data; Travel time reliability
• ISSN: 0968-090X; 1879-2359
• DOI: 10.1016/j.trc.2019.09.014

•A framework for modeling the effect of perceived reliability on mode choice.•Cumulative Prospect Theory-based model using observational data.•Estimation of the value and the weighting function parameters from observed data.•Discussion on the extensions of the model to other choice dimensions. This paper utilizes the well-known Cumulative Prospect Theory (CPT) to study how travel mode choice is affected by travel time reliability. Prospect Theory (PT) was originally developed to model decision making under risk, and later extended to CPT, which is also applicable to decision making under uncertainty. Travel time reliability is related to uncertainties in travel time, and its effect on travel behavior can be properly modeled using CPT. PT and CPT applications in transportation are mainly limited to experimental studies. This paper introduces a CPT-based framework for mode choice modeling using observational data. The framework can help travel demand modelers utilize CPT for modeling reliability in their trip-based or activity-based models, which are typically estimated from observational data (household travel surveys). To showcase model applications, the model parameters are estimated using a combination of revealed preference household travel survey data and empirically observed travel time data for a mode choice scenario in the Washington, D.C. area. We estimated the parameters of the utility function, the parameters of the CPT value function, and the parameters of the CPT weighting function and discussed how they imply diminishing sensitivity, risk-aversion in small probability losses, and risk-seeking in high-probability losses. The paper is focused on mode choice, but its extension to other choice dimensions is discussed.