City-Level Agent-Based Multi-modal Modeling of Transportation Networks: Model Development and Preliminary Testing

Digital cities have the potential to produce major improvements in the transportation system with the eminent availability of data and the use of advanced modeling techniques. This modeling entails a number of challenges, namely: the network scale that covers large urban multi-modal transportation networks and the trade-off between model scalability and accuracy. This paper introduces a novel simulation framework that efficiently supports large-scale agent-based multimodal transportation system modeling. We call this framework “INTEGRATION Ver. 3.0”, or “INTGRAT3” for short. The INTGRAT3 framework utilizes both microscopic and mesoscopic modeling techniques to take advantage of the strengths of each modeling approach. In order to increase the model scalability, decrease the model complexity, and achieve a reasonable simulation speed, the INTGRAT3 framework utilizes parallel simulation through two partitioning techniques: spatial partitioning by separating the network geographically and vertical partitioning by separating the network by transportation mode for odes that interact minimally. The INTGRAT3 framework creates multimodal plans for a portion of the trips (controlled trips) and tracks the traveler’s trips on a second-by-second basis across the different modes. We instantiate this framework in a system model of Los Angeles (LA) supporting our study of the impact on transportation decisions over a 6 h period of the morning commute (6am–12pm). The results show that by modifying travel choices of only 5% of the trips, large reductions in traffic congestion (ranging from 3.8 to 17.2% reductions in vehicle delay) are achievable with marginal reductions in vehicle fuel consumption levels (ranging from 1.4 to 3.5%).