Using optimization methods can bring forth additional functionalities which are not possible with conventional methods. In principle, an optimized solution may perform suboptimaly when errors in implementation or model parameters are encountered. I focus my research on the real-world applicability of the solutions.
The goal of my research is to robustly solve and optimize problems in environments affected by errors in implementation, in modeling, and in parameter specification amongst others. One of my research projects focuses on using nanoHUB data to develop robust methods for classification of users and resources and for prediction of policy outcomes. After taking data uncertainty into account, solutions are more reliable making evaluation processes more efficient and decision-making processes more effective.