Data-Driven Methods in Simulation and Optimization Modeling
Stochastic simulation modeling is becoming more widely used to model real-world complex systems and optimize the system performance. Traditional approach, however, aims to draw insights based on precisely known model assumptions. When the models are unknown and are only observed via data, the complexity grows and the drawn insights can contain significant errors. This project investigates the remedies of both issues in simulation modeling and stochastic optimization problems. One type of schemes that we will study is to reweight the empirical data that represents the unknown model in ways that can quantify the errors analytically. We will analyze the procedures and the tractability of these schemes for large-scale and dynamic problems where the complexities can be challenging. The research will have applications in areas including transportation, energy, and operations management, where stochastic and simulation modeling is used and data is routinely collected.
Presented at INFORMS, Houston, TX.
Presented at the SIAM Conference on Uncertainty Quantification, Garden Grove, CA.
Presented at IISE Annual Conference & Expo (ISERC), Pittsburg, PA.
(Re-scaled) Multi-Attempt Approximation of Choice Model and Its Application to Assortment Optimization for publication
Published article in Production and Operations Management
Published article in Methodology and Computing in Applied Probability.
$274,999 grant from the National Science Foundation.
Presented in Anaheim, CA.
Published in IEEE Transactions on Power Systems.
Published in IISE Transactions..
When Wind Travels Through Turbines: A New Statistical Approach for Characterizing Heterogeneous Wake Effects
Presented in Auburn, Alabama.
When wind travels through turbines: a new statistical approach for characterizing heterogeneous wake effects in multi-turbine wind farms
Published in IISE Transactions.
Presented in Nashville, TN.
Presented in Pittsburgh, PA.
High-Fidelity, High-Performance Multi-Stage Transmission Planning with Spatio-Temporal Uncertainty Models
$ 431,175 from the National Science Foundation.