Holistic, integrated generation of daily-activity plans for Switzerland: from population synthesis to trip generation

Abstract

Detailed and realistic input travel demand are required in agent-based transportation simulators in order to generate high-fidelity outputs. However, most prior work has focused separately on either population synthesis, the process of travel demand generation, or modal choice models. This paper presents a holistic and integrated approach that integrates the different elements – population synthesis, job assignment, car ownership and mode choice models, and the trip generation procedure – in order to generate a realistic synthetic population and microscopic daily-activity plans for the large-scale scenario for the entire population of Switzerland. The generated Swiss travel demand (about 5 million agents) is then executed for the entire transportation network of Switzerland using our in-house developed, GPU-accelerated mobility simulator, GEMSim, which is integrated within our agent-based simulation framework, EnerPol, to simultaneously address the impacts of transformations in mobility on energy infrastructure, land use and urban development. The insights from the simulated scenarios, and the needs for what additional data may be useful to gather are presented. It is also demonstrated how the outcomes from the high-resolution framework can be used by policy makers and planners at the city, cantonal, regional and federal levels.