Random Utility Maximization model considering the information search process

Abstract

Choice modelling has been dominated by static representations preferences due to their ease of implementation, transparent economic interpretability, and statistical coherency. Unlike, the Decision Field Theory (DFT) model explicitly includes the attribute scrutiny process within the choice decision, making it more closely related to the behavior that is observed in practice. However, the DFT model lacks the RUM model's microeconomic interpretability and has statistical limitations regarding the identification of the model parameters. This research introduces the "RUM-DFT" model, encompassing ideas from both approaches. Using Monte Carlo simulations and applying the proposed model to a database of real choices, it is first shown that the proposed approach can properly identify the parameters of the deliberation process, replicate the dynamic behavior of the utilities during the deliberation process; and retains full economic interpretability, since the estimated coefficients correspond to marginal indirect utilities when there is perfect knowledge of the information search process.