Location choice in the context of multi-day activity-travel patterns: model development and empirical results

Multi-day activity-based models of travel demand are receiving increasing interest recently as successors of existing single-day activity-based models. In this article, we argue that predicting activity location choice-sets can no longer be ignored when multi-day time frames are adopted in these models. We develop a model to predict activity location choice-sets and choices from these sets conditionally upon varying activity schedule contexts. We propose a method to estimate parameters of the involved utility functions that do not require observations or imputation of choice-sets. This is achieved by using Bayes’ method to transform the likelihood of chosen locations into a likelihood of attribute profiles of chosen locations. An application of the method using a national travel diary dataset illustrates the approach.