Abstract
One of the advanced methods of modeling activity duration is the multiple discrete-continuous extreme value (MDCEV) model. The translating satiation parameter of the model is intended to capture the constant marginal utility effect, but model testing shows that it does not do so. In this paper, the structure of the translating satiation parameter was modified to incorporate an additional parameter that enables the model to capture this effect. Two datasets from cities in Saudi Arabia and Germany were used to test the applicability of the modified model for any dataset. The results showed that the proposed model structure increased the accuracy of activity duration prediction by up to 74%. The modified model represents an improvement to the conventional MDCEV model in terms of accuracy and flexibility, and as such should be a valuable tool in transportation planning and management.