When transit systems undergo change, such as when new routes are added or schedules are updated, it can impact mobility, energy use and emissions throughout the entire transportation system in complex and often unexpected ways.

The U.S. Department of Energy’s (DOE) Argonne National Laboratory has been helping the City of Chicago to better predict those effects using high performance computing capabilities.

Researchers are partnering with the Chicago Transit Authority (CTA) to develop a new approach for understanding how transportation systems and people are likely to respond to systemic changes. This approach pairs algorithms that are designed to optimize route designs, timetables and strategies for better performance with POLARIS, Argonne’s transportation system simulation tool.

Simulating system response with POLARIS

Transportation agencies like CTA are regularly considering a large number of changes to increase ridership and access to services, as well as curb emissions and energy use. Measures could include adding new routes, shortening or lengthening existing ones, changing timetables or implementing new technologies like electric buses.

With any of these changes, agency officials must weigh their effect on the remaining parts of the system. To model these responses, agencies have historically relied on elasticity functions or statistical models but these offer limited insights.

For example, if planners were considering adding stops to a bus route, statistical models would shed light on the number of riders they’d be likely to gain. However, they could not explain how people’ choice to ride the bus might affect other decisions they make throughout the day that impact their mobility.

To capture these downstream and system level effects, Argonne researchers are modeling traveler responses using POLARIS by simulating the decision-making and subsequent movement of millions of agents throughout an entire day.

“By considering the impact of transit on the overall transportation system from an agent-based perspective, we are able to provide unique insights, including the secondary and tertiary effects,” said Omer Verbas, Argonne computational transportation engineer.

Coupling POLARIS with optimization algorithms

Researchers are coupling POLARIS with different optimization algorithms, including ones designed to optimize route design, frequency and route electrification.

These algorithms output such as new routes are then emedded into POLARIS by the researchers.  POLARIS simulates the system response, including how many people will board each and every bus or subway and the times that each bus in Chicago will arrive and depart.

Based on the simulation outputs, the algorithm improves itself, creating new sets of timetables, frequencies, modified designs and so on, and researchers iterate the process hundreds of times for different scenarios until their solutions stabilize.

“With these new capabilities, the CTA and other transit agencies will have resources to make more informed decisions about the future of transportation within our cities,” said Josh Auld, Argonne Technical Manager for Transportation Systems and Mobility.

This project is part of HPC4Mobility, a program within the U.S. Department of Energy’s (DOE) HPC4 Energy Innovation initiative. HPC4Mobility facilitates access to high-performance computing (HPC) capabilities and expertise across the DOE National Laboratories to deliver solutions that could revolutionize mobility and transportation.