Jointly Modelling Individual's Daily Activity-travel Time Use and Mode Share by a Nested Multivariate Tobit Model System

Understanding mechanisms underlie the individual's daily time allocations is very important to understand the variability of individual's time-space constraints and to forecast his/her daily activity participation. At most of previous studies, activity time allocation was viewed as allocating a continuous quantity (daily time budget) into multiple discrete alternatives (i.e. various activities and trips to engage with). However, few researches considered the influence of travel time that needs to be spent on reaching the activity location. Moreover, travel time itself is influenced by individuals’ mode choice. This can lead to an over- or under-estimation of particular activity time location. In order to explicitly include the individual's travel time and mode choice considerations in activity time allocation modelling, in this study, a nested multivariate Tobit model is proposed. This proposed model can handle: 1. Corner solution problem (i.e. the present of substantial amount of zero observations); 2. Time allocation trade-offs among different types of activities (which tends to be ignored in previous studies); 3. Travel is treated as a derived demand of activity participation (i.e. travel time and mode share are automatically censored, and are not estimated, if corresponding activity duration is censored). The model is applied on a combined dataset of Swedish national travel survey (NTS) and SMHI (Swedish Meteorological and Hydrological Institute) weather record. Individuals’ work and non-work activity durations, travel time and mode shares are jointly modelled as dependent variables. The influences of time-location characteristics, individual and household socio demographics and weather characteristics on each dependent variable are examined. The estimation results show a strong work and non-work activity time trade-offs due to the individual's time-space constraints. Evidences on a potential positive utility of travel time added on non-work activity time allocation in the Swedish case, are also found. Meanwhile, the results also show a consistent mode choice preference for a given individual. The estimated nested multivariate Tobit model provides a superior prediction, in terms of the deviation of the predicted value against the actual value conditional on the correct prediction regarding censored and non-censored, compared to mutually independent Tobit models. However, the nested multivariate Tobit model does not necessarily have a better prediction for model components regarding non-work related activities.