Research performed in CADO combine mathematics, statistics, optimization and programming to transform large datasets into usable knowledge to help make intelligent decisions. Our goal is to advance the education and research in analytics based solution to real-world problems.
Center for Applied Data Sciences and Optimization
CADO is a multi-disciplinary laboratory where students conduct research and apply their knowledge to support decision making in diverse settings such as manufacturing, logistics, transportation, and healthcare.
Current projects include:
- Using drones for emergency medicine delivery
- Data analysis and optimization of US heart allocation system
- Interactive maps for data visualization of organ transplants in the US
- Network algorithm design for Starboard Corp
- Logistics optimization for a crude Oil Company
- Network analysis of Illicit organ Sales networks
- Machine learning algorithms for predicting pavement conditions
- Optimization approach to workload balancing to combat physician burnout in a healthcare network
- Data driven approach for estimating residential energy consumption
- Optimal network design for electric vehicle charging infrastructure
- Multi-level optimization models and algorithms for demand response in smart grid
"Operations research attracted me early in my career, as a perfect tool to put mathematical rigor to practical use in so many diverse applications. To this day, I continued to be challenged by, and enjoy applying analytics and optimization to provide solutions in industry and academic settings."Lihui Bai, Co-Director
Kidney Paired Donor (KPD) exchange is the system that allows incompatible pairs to exchange donors with other incompatible pairs to improve donor–recipient compatibilities. The integration of KPD with desensitization therapy can increase the number of people who are able to exchange donor. We apply an optimization simulation approach to evaluate the impact of optimally integrating a desensitization protocol in a KPD program.
Attempting to increase energy efficiency and improve system load factors in an electricity distribution system, demand response (DR) has long been proposed and implemented as a form of load management. Various pricing structures incentivizing consumers to shift energy consumption from on-peak to off-peak periods are evident in this field. Most DR methods currently used in practice belong to static variable pricing (e.g., Time of Use, Critical Peak Pricing) and the impact of such tariffs has been well established. However, dynamic variable pricing in general is less studied and much less practiced in the field, due to the lack of understanding of consumer behavior in response to price uncertainty. We study a novel dynamic variable pricing scheme that uses the coincident demand charge to reduce load consumption during peak events. Multiple techniques including multi-attribute utility function, model predictive control, and conditional Markov chain are used to model consumer behavior and thus predict the system peak. Effects of various residential electricity rates based on coincident demand charge are studied, under an integrated electricity distribution system with renewable solar production.
FEATURED Faculty: Monica Gentili, Co-Director
Each day 20 people die waiting for transplants; every 10 minutes another person is added to the waiting list. The gap between the number of organs donated and the number of organs in need continues to grow. More lives can be saved by improving the current organ allocation system.