Activity-based travel demand generation for digital twins in large-scale microscopic traffic simulation

Rapid advances in transport technologies and daily travel-demand patterns are reshaping why, where, when, and how people travel. Remote work, flexible study schedules, and e-commerce are weakening traditional peak-hour dynamics, dispersing trips and increasing temporal variability, while open-source data enables richer representations of transport systems. These shifts call for next-generation models that integrate multi-source data and test policy or technology “what-if” scenarios within city digital twins. Although activity-based models (ABMs) can represent complete activity–travel sequences at the individual level, large-scale microscopic ABMs still often depend on costly disaggregated travel-diary surveys, and 24-hour microscopic simulations remain uncommon. This study proposes a transferable framework for building daily activity-based travel-demand models from readily available basic mobility data and other easy-to-collect big-data sources. First, an existing peak-hour ABM is improved and re-validated to assess “superblocks,” a sustainable traffic intervention, using microscopic simulation. The evaluation quantifies changes in modal split, network performance and individual travel times. The main contribution of this study is the development and comparison of two cost-effective daily ABM approaches suitable for large-scale microscopic simulation. The first reconstructs daily demand from peak-hour O–D matrices and traffic-detector counts by scaling demand and spreading departures across the day, producing temporally and spatially detailed primary chains (home–work–home). The second generates daily activity–travel diaries by combining aggregated population and mobility statistics combining open-source land-use big data; choose compatible facilities subject to opening-hour, capacity, and daily travel-time-budget constraints, enabling linked primary and secondary trips and demographic heterogeneity. Generated plans feed 24-hour microscopic simulations in HybridPy/SUMO for a case study in Bologna, Italy. Validation shows moderate but systematic correspondence between simulated and observed flows, and daily travel times and activity durations are analysed by mode, age, and gender. Future work calls for advances in multimodal routing speed, household interaction modelling, and computational efficiency.