A two-stage recreational destination choice study incorporating fuzzy logic in discrete choice modelling

Abstract

A major challenge in developing discrete choice models when the feasible choice set is extensively large is to find an appropriate choice set that is computationally manageable for being integrated into the choice modelling specification while it also reflects the behaviour of the decision-maker. This study aims to develop a new sampling method for choice set formation which is then used to improve the knowledge about recreational destination choice behaviour. The study proposes a two-stage framework where in the first stage a rule-based fuzzy logic model is used to form a latent choice set. The second stage incorporates the set in a discrete choice model to estimate destination selection behaviour. In the travel demand modelling literature, there has been little research into modelling recreational activity destination choices. In this area, the primary focus has been on other activities, such as work, study and shopping. In this context, the sampling methods used to form the choice set using the large set of possible options in the universal set are unconvincing. The proposed two-stage approach is examined using a household travel survey from Victoria, Australia, where nine different recreational destination choice models are developed. Compared to the conventional modelling framework, these models show impressive model fit and prediction potential.