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

Srushti Rath’s MS thesis work on air taxis planning suggests access to airports as an initial market that can help drive investment. We developed a hub location model that is sensitive to mode choice between ground taxis and the air taxis. Findings suggest that 9 skyports would be adequate to serve such demand to accommodate variations in transfer times from ground taxi to the skyports. The research was partially supported by the C2SMART Center under USDOT grant #69A3551747124.

The paper can be found here: https://www.sciencedirect.com/science/article/abs/pii/S0969699722001132

Given the importance of data sharing between mobility providers, it is a highly understudied topic area. Qi Liu studies the properties of data sharing between competitive mobility (transit) providers as a coopetitive game with coalition structure formation for partitioning data sharing and Bayesian game to compute the value of noncooperative equilibria under such a setting. We highlight that having everyone share with each other is not a foregone conclusion, even under complementary services (e.g. multimodal trips). The insights from this work can help guide policy toward more incentive-aware data sharing structures, quantifying subsidies or revenue allocations that can encourage cooperation, and for mobility aggregators (e.g. Whim, Transit App, Cubic, and the like) to consider data sharing requirements between different members.

The work was funded by NSF CMMI-1652735 and USDOT #69A3551747124.

Link to paper: https://www.sciencedirect.com/science/article/pii/S0191261522001126

For mobility companies interested in global city classification data to help planning deployments, this work by Srushti Rath crowdsourced original data from Wikipedia using a new natural language processing algorithm. We expanded city classifications developed by Jimi Oke for ~300 cities up to a set of ~2000+ global cities. Links to data and code are included in the paper. Developed as part of a collaborative C2SMART project with Via.

Paper link: https://www.sciencedirect.com/science/article/pii/S0968090X22001048?dgcid=author

Qi Liu developed an algorithm to estimate transit passenger flows at a network level (the first of its kind) using stop count data (e.g. Transit Wireless, smartcard data, etc.). Testing on Shanghai bus data in Qingpu District with 4 bus lines and 120 segments, we show the algorithm leads to average of the segment flows that are only 2.6% off from the average of the observed flows. Code is available on Github and should be of interest to transit operators (e.g. NYC Transit Authority). Funding support from C2SMART and U.S. National Science Foundation (NSF) CMMI-1652735.

Paper link: https://www.tandfonline.com/doi/abs/10.1080/21680566.2022.2060370

In this collaborative work with UNSW, we present a modified dial-a-ride problem that incorporates user demand response, and several heuristic algorithms to solve it. Such a model can evaluate different pricing structures (tested zone-based, distance-based, and flat fee structures using data from NYC, and found that a zonal scheme can provide the best revenue and ridership optimization). This can allow planners to design service zones with different pricing structures (including integrated fare bundles with transit in a MaaS environment), integrate pricing with route and destination recommendation, and equity considerations with customized pricing for underserved population segments.

The paper can be found here:
https://www.sciencedirect.com/science/article/pii/S1366554521003562?dgcid=coauthor

From an NYU Vertically Integrated Projects (VIP) collaboration with NYC DCAS, our lab developed a tool to analyze a population’s access to different charging station location configurations and station mixes considering queue delays (which we used to analyze Revel’s charging hub last year in a Linkedin article). In this publication, we present the underlying algorithm and evaluate DCAS investment policies of adding DCFC chargers to existing stations. A policy of adding DCFC to stations based on those with highest utilization ratio was more effective than choosing those with highest queue delays.

The paper can be found here:
https://www.tandfonline.com/doi/full/10.1080/15568318.2022.2029633