Projects

Smart and Equitable Public Transit

I am one of the lead investigators of SmartTransit.AI, a multidisciplinary research team comprising computer scientists, civil engineers, social scientists, urban planners, and public transportation experts dedicated to designing innovative solutions for enhancing public transportation operations. Our focus is on improving availability, reliability, effectiveness, and efficiency. Funded through federal grants and in collaboration with partner agencies such as the Chattanooga Area Regional Transportation Authority (CARTA) and Nashville WeGo, we design AI-based algorithms to address integrated multi-modal logistics challenges at scale, incorporating both same-day and long-term future trends. A key aspect of our work is the design of models for real-time energy consumption of mixed-vehicle fleets, including electric, hybrid, and diesel vehicles. 

Funded by the National Science Foundation (PI), US Department of Energy (Co-PI), and the state of Tennessee TNGo (Co-PI).

Key Publications

ADVISER: Resource Allocation for Improving Maternal and Infant Health

I was the PI for the "ADVISER: AI-driven Vaccination Intervention Optimizer." Over the last three years, I have collaborated with HelpMum, a non-profit agency based in Nigeria, and Google AI for Social Good to optimize the allocation of health interventions under limited resources. ADVISER is based on designing machine learning models to forecast immunization outcomes and algorithmic approaches for solving large-scale mathematical programs for resource allocation. ADVISER is currently deployed in thirteen local governments in Nigeria, improving immunization outcomes by over 30%. 

Funded by Google AI for Social Good (PI) and the Patrick J McGovern Foundation (PI).

Key Publications

Robust and Adaptive Decision-Making for Non-Stationary CPS

A major thrust of my research involves designing adaptive algorithms for decision-making under uncertainty. In many real-world applications, agents must make sequential decisions in environments where conditions are subject to change due to various exogenous factors. These non-stationary environments pose significant challenges to traditional decision-making models, which typically assume stationary dynamics. Non-stationary Markov decision processes (NS-MDPs) offer a framework to model and solve decision problems under such changing conditions. I work on designing approaches to detect such changes and then adapt to them online.


Funded by the National Science Foundation (Co-Investigator) and DARPA.


Key Publications

Multi-Agent Systems for Emergency Response

Planning and preparation in anticipation of urban emergency incidents are critical because of the alarming extent of the damage such incidents cause and the sheer frequency of their occurrence. From the perspective of stochastic control processes, proactive incident stationing and response are semi-Markovian decision processes that are computationally difficult to solve. Over the past six years, my research on algorithmic approaches to ERS has developed proactive stationing and principled dispatch strategies to reduce overall response times. 

Funded by the National Science Foundation (lead student during PhD) and Tennessee Department of Transportation (Co-PI).


Key Publications

Fair and Interpretable Decision-Making

As decision-making approaches increasingly use large volumes of data and complex neural networks, we must design equitable and fair approaches. At the same time, we must design interpretable algorithms to ensure that domain experts and end-users can trust and understand how algorithms optimize decisions. While much prior work has focused on exploring the fairness and explainability of learning-based approaches, decision-making has remained largely unexplored. I work to design algorithmic approaches for sequential decision-making and combinatorial optimization that are fair and explainable.

Key Publications