Building on our earlier work, we further expand the classic static dial a ride problem to model platooning between demand-responsive modular vehicles in serving customers. We invented an efficient heuristic and showed using a network of Anaheim CA how platooning can help reduce vehicle travel cost and passenger service time. Statistical tests with computational data suggest such operations and technology benefiting most from serving “enclaves” of passengers. This work should be of interest to microtransit/ridepooling or delivery operators considering more flexible vehicular technologies.

The research was funded by NSF CMMI-2022967 and USDOT #69A3551747124 via the C2SMART Center. It was also presented at the 15th International Conference on Advanced Systems in Public Transport at Tel Aviv in 2022.

Paper link: https://www.sciencedirect.com/science/article/abs/pii/S0968090X23001808

In this collaborative work with TU Delft, we invented a new insurance policy for mobility providers that allows them to borrow fleet resources from each other in the event of a major disruption. Using disruption data from transit operators in The Netherlands, we show how different sized operators carry different amounts of negotiating power in such risk pooling agreements, and how this depends on the underlying disruption risks. Should encourage more collaborative resiliency discussions between different mobility providers.

The research was funded by NSF CMMI-1634973 and CriticalMaaS no. 804469, as well as support from USDOT 69A3551747124 via the C2SMART Center.

Link to the paper: https://www.tandfonline.com/doi/abs/10.1080/23249935.2023.2210229

Whereas conventional transportation planning tends to be region-centric, we see a shift now to one that is more operator-centric, as transport providers work with different city and regional agencies. As such, understanding service impacts on different communities are more important than ever, particularly for emerging technologies. Microtransit/ridepooling is one such technology, and the limited observations (e.g. dozens) we have alone might not provide a robust statistical picture for understanding deployment needs across a landscape of different market typologies for thousands of cities. Our C2SMART Center study, in collaboration with Via, rectifies this by proposing the use of simulation-based “scenario data upscaling” (in much the same way algorithms can upscale images) to inform on metrics needed to estimate forecast models for managing deployment portfolios. Even with data from only a handful of cities, we were able to generate meaningful forecast models based on hundreds of synthesized scenarios, showing for microtransit deployment how to evaluate different service regions across a portfolio of different cities in the U.S. Should be of interest to federal policymakers, microtransit/ridepooling providers, and other emerging transportation technology providers and local agencies, particularly as we look to revamp the national tools and data we use for planning transport infrastructure.

Paper available here: https://www.sciencedirect.com/science/article/pii/S0965856423000046

With emergency vehicle response times increasing due to increasing congestion, there needs to be more technological solutions beyond using only sirens. We show how using reinforcement learning for route guidance integrated with signal control can reduce EMV response time by up to 43% in simulated scenarios over existing technologies. Should be of interest for discussion between traffic agencies (New York City Department of Transportation) and EMVs (New York City Fire Department, NYU Langone Health). Part of Haoran Su’s dissertation and in collaboration with Siemens. Funding from Siemens, Dwight David Eisenhower Transportation Fellowship, C2SMART University Transportation Center, NSFC Project 62103260, SJTU UM Joint Institute, J. Wu & J. Sun Endowment Fund, and the NYU University Research Challenge Fund 2020.

Paper can be found here: https://www.sciencedirect.com/science/article/pii/S0968090X22003680

Xiyuan Ren considered a ubiquitous data setting where we can fit choice models at the individual level (using inverse optimization-based machine learning), and showed how this can better capture joint travel choices within an activity scheduling context. We demonstrated this with data from Shanghai, where the agent level mixed logit model (fitted individual coefficients) outperforms state of the art dynamic discrete choice models. We also show that this can better integrate with optimization models to support choice-based system design. The work is now being applied to our C2SMART Center project in collaboration with Replica to estimate constrained mode choice models across population segments from their synthetic population data for New York State to design equitable statewide mobility services.

The research was funded by NSF CMMI-1652735 and the C2SMART Center (USDOT #69A3551747124).
Paper here: https://www.sciencedirect.com/science/article/pii/S0191261522001862

A new algorithm was developed to schedule both traditional and next-gen quay cranes at container terminals. It allows us to evaluate different levels of next gen quay crane investments for the Abu Dhabi Ports, which can reduce makespan by as much as 70% when all traditional cranes are replaced. Part of Omar Abou Kasm’s PhD dissertation work.

Paper: https://www.sciencedirect.com/science/article/pii/S0377221722008232