Random Utility Maximisation Model Considering the Information Search Process

Abstract

Decision-making is a subject of interest in various scientific fields, including transportation. Choice modelling, which aims to understand and predict decision-makers' behaviour, has predominantly relied on static representations of preferences, principally through the Random Utility Maximisation (RUM) model. The RUM model has been widely used due to its ease of implementation, economic interpretability, and statistical coherence. However, the RUM model assumes that decision-makers possess complete information about all alternative’s attributes and can instantaneously process and recall it, which may not align with actual human behaviour.In contrast, the Decision Field Theory (DFT) model explicitly incorporates the attribute scrutiny and recalling process within the decision-making framework, making it more realistic. However, the DFT model lacks microeconomic interpretability and faces statistical challenges in parameter identification.To bridge this gap, the "RUM-DFT" model is introduced, combining aspects from both approaches. Through Monte Carlo simulations, the proposed model is shown to: i) allow the identification of parameters related to the deliberation process, ii) replicate the dynamic behaviour of utilities during deliberation as observed in practice, iii) maintain economic interpretability by estimating coefficients that correspond to marginal indirect utilities under full information conditions, and iv) highlight the pitfalls of using a RUM model that disregards the true dynamics of data generation process, leading to inconsistent estimators and potentially flawed policy recommendations.The SwissMetro database is used to test the feasibility of the RUM-DFT model. A version of the model assuming homogeneous maximum deliberation time is implemented, demonstrating better goodness-of-fit measures compared to conventional choice modelling approaches, including the DFT model. These promising results encourage further research, particularly regarding the numerical challenges arising from a growing or unknown number of scrutiny steps faced by decision-makers.

Keywords: Dynamic-cognitive approach, RUM, Decision field theory, Information search process, Choice modelling