ABM3 Model Development

The San Diego Association of Governments (SANDAG) is responsible for regional transportation planning and aims to create a sustainable, efficient, and equitable transportation system for the San Diego region and its residents. To that end, SANDAG commissioned the development of a next-generation activity-based model, ABM3, to gain deeper insights into transportation and land-use projects, thereby ensuring alignment with broader sustainability, equity, and resilience goals. Notably, the ABM3 model leverages recent data from household travel surveys, providing a comprehensive framework for forecasting demand, measuring the impacts of new technologies, and supporting SANDAG’s long-term planning objectives.

The Challenge

SANDAG’s previous ABM2+ model was increasingly limited in capturing evolving travel patterns and emerging technologies. For example, shifts like increased telecommuting, the rising popularity of ride-hailing, and the integration of micromobility options highlighted the need for a model that could more precisely reflect San Diego’s changing travel dynamics. Furthermore, with remote work growing significantly due to the COVID-19 pandemic, SANDAG recognized the importance of recalibrating its models to better represent these new behaviors.

In addition, the agency also faced challenges in incorporating future-focused scenarios, such as the adoption of autonomous vehicles and the use of managed lanes. Consequently, meeting these needs required integrating high-resolution, real-world travel data. This data would capture emerging trends and, equally important, provide the flexibility and transparency needed to comply with state and federal planning requirements.

RSG's Solution

To address these complex challenges, RSG collaborated with SANDAG on travel model development for the ABM3 model, using ActivitySim, the leading open-source platform for activity-based modeling. A key innovation in this project was the integration of real-world household travel survey data collected using rMove™, RSG’s location-aware market research platform that supports multi-day travel diaries. This data collection proved instrumental in capturing shifts in travel behavior, particularly the marked increase in remote work and telecommuting behaviors. Together, these features collectively improved the model's ability to reflect both current and future travel patterns.

In terms of the model’s technical structure, RSG introduced disaggregate accessibilities, which replaced the simpler aggregate method used in ABM2+. As a result, the ABM3 model can more accurately simulate travel costs and times across different modes, thereby improving policy analysis and enabling more granular forecasting capabilities. Additionally, other notable enhancements include replacing the shadow pricing mechanism with a more efficient algorithm and making advancements in modeling autonomous vehicle ownership and vehicle type choice. Specifically, the enhanced vehicle choice model allows SANDAG to assess the impact of electric vehicles as they increasingly become part of the vehicle fleet over time. Altogether, these enhancements provide SANDAG with a more nuanced understanding of household travel behaviors.

Overall, ABM3 is a foundational tool for SANDAG’s 2025 Regional Plan and future initiatives. By integrating high-resolution, real-world household travel data with advanced modeling techniques, ABM3 delivers a nuanced, data-driven understanding of the San Diego region’s travel patterns and evolving transportation needs. Furthermore, this model empowers SANDAG to evaluate the impacts of policy changes, test sustainability initiatives, and explore equity-driven planning scenarios with confidence. Ultimately, this ensures that transportation investments are data-driven and aligned with the region’s goals. Looking ahead, ABM3 will allow SANDAG to stay at the forefront of innovation, supporting its mission to create a sustainable, efficient, and equitable transportation network for the San Diego region and its residents.