Numerical estimation of travel mode preferences for agents in agent-based simulations

Passenger flows predictions or forecasting are a critical element in the planning, dimensioning, and operating of transport systems. Agent-based simulation models are the first models that forecast travel behavior based on a mix of data that includes both behavior patterns and daily schedules of individual users in the prediction algorithms, plus an element of randomness. Nevertheless, agent-based simulation models only use quantitative data to design the profile of the agents. Researchers agree that qualitative, experiential, and consultative methods are critical to transport and urban planning. Therefore, we propose a simple methodology that can support a numerical estimation of the preference for different modal choices of an archetypal user. The estimation is done based on existing travel behavior data and can provide input for updating agent profiles used in agent-based simulations in the field of passenger transport. The results of our study suggest that the proposed methodology is accurate in estimating the travel mode preference of predefined persona profiles and translating it into a numerical format. Our findings set a solid base for bridging persona generation models, based on a mix of quantitative and qualitative data, with defining agent profiles and rules for agent-based simulations for the field of passenger transport.