A behaviorally sound activity-based travel demand model would enable a realistic analysis of emerging mobility options and their integration into the complex transportation system. Available regional travel demand models in the Greater Toronto and Hamilton Area (GTHA) use aggregate, static, and deterministic user equilibrium-based network models (e.g., the GTA and GGH models). No operational demand model is available of the entire GTHA’s multimodal transportation network that can capture the network dynamics and agent-based network microsimulation.
This paper presents such a model. It is based on integrating an activity-based modelling framework named CUSTOM (Comprehensive Utility Maximizing Travel Options Modelling) with a network model named GTASim that was developed using a multi-agent-based transport simulation framework to cover the entire GTHA region.
The CUSTOM model considers a 24-hour modelling time frame for the activity-travel scheduling process. It is based on the random utility maximization theory and provides individuals’ joint choice of activity type, location, departure time and mode. The mode choice component is implicitly history-dependent through the choice set formulation and thus, captures the most fundamental physical constraints (i.e., tour constraints). The CUSTOM framework generates preferred activity-travel plans for agents representing the demand of the urban transport system. The GTASim model executes the plans assigning travellers to the road and transit networks representing the supply. The outcome of the demand-supply interaction is the urban transport system’s level of service (LOS). The LOS is then fed back to the scheduler, generating adjusted activity-travel plans. This feedback mechanism represents the reality that trip markers modify their behaviours considering scarce and sometimes costly transport network supply.
However, the GTASim and CUSTOM model is calibrated independently to the Transportation Tomorrow Survey 2016 household travel survey dataset, and the feedback loop is still missing. Moreover, the current GTASim cannot accurately account for the time-varying monetary cost of travel. Hence, this paper makes two substantial and noteworthy contributions. Firstly, the study extends GTASIM’s capacity to account for the monetary cost of individual trips in the scoring of each trip leg (with a focus on transit fares). Then, developing the integrated framework which makes it transferable to any study context. The GTASim model with the new capability is further compared to the available EMME-based regional model.
Our adaptable and forward-thinking framework considers changing travel preferences and provides authorities with a powerful tool for meeting future transportation demands while ensuring safety and efficiency for all.
