Skip to content
Ran Chen headshot

Ran Chen

Ran Chen

Postdoc Associate at Massachusetts Institute of Technology

Dr. Ran Chen is currently a postdoc associate at Laboratory for Information and  decision Systems at MIT. Her research focuses on combining statistics, optimization, and machine learning to address challenges motivated by social science (including business) and health care. She completed her Ph.D. in Statistics and Data Science from the Wharton School, University of Pennsylvania. She obtained B.S. in pure and applied mathematics from Tsinghua Xuetang Mathematics Program at Tsinghua University.

Track: Revenue Management

High-dimensional Continuum Armed and High-dimensional Contextual Bandit: With Applications to Assortment and Pricing

Both assortment and pricing are important problems in operations research, marketing, and revenue management. While these two problems usually appear together, current literature addresses them separately. We formulate the joint assortment and pricing problem as a high-dimensional continuum armed and high-dimensional contextual bandit problem. Our formulation is structure-rich and interpretable where demand function is implicitly incorporated. Recent developments in contextual bandit problems focus on settings where the number of arms is small, hence impracticable with high-dimensional continuous arm spaces. We propose an efficient bandit algorithm for our new model with theoretical justification. We demonstrate the effectiveness of our algorithm to jointly optimize assortment and pricing for revenue maximization for a giant online retailer. In addition, the generality of our model makes several bandit problems its special cases and allows wider applications in business and healthcare. Simulation studies show our algorithm’s superiority over mainstream bandit algorithms in their applicable settings.