This paper presents an activity-based transport demand modelling framework for regional contexts, where inter-regional commuting is important and behavioural data are often sparse. Using Greater Bendigo, a regional city in Australia, as a case study, we develop daily activity chains and subsequent activity-based demand for use in agent-based simulation models. The methodology integrates statewide travel flows with detailed local modelling through demographic synthesis, activity-chain generation, spatial location assignment, and mode-choice modelling.
The key contribution is a multi-scale location assignment framework that prioritises the mandatory activities, work and education, before allocating non-mandatory activities. Work locations are assigned to reproduce observed inter-regional commuting flows and distance distributions derived from Journey to Work (JTW) data, while education locations are allocated subject to facility-level capacity constraints. Secondary activity locations are assigned using a two-sided gravity-based approach conditioned on preceding and subsequent activity anchors, land-use-based destination attractiveness, and empirically observed mode-specific distance distributions. Trips entering and leaving the local study region are explicitly incorporated to maintain internal behavioural consistency.
The activity-based demand is assessed against observed population totals, activity timing, triplength distributions, commuting patterns, and regional mode shares. The comparison results show close agreement with key calibration benchmarks and plausible agreement with complementary survey-based checks. The generated demand includes about 13% cross-boundary work trips, underscoring the importance of explicitly accounting for external travel in regional demand models. Overall, the proposed methodology provides a scalable and transferable approach for generating activity-based demand for regional cities while preserving cross-boundary travel consistency.