Paat Rusmevichientong is the Justin Dart Professor of Operations Management and Professor of Data Sciences and Operations in the Marshall School of Business at the University of Southern California. Prior to joining the Marshall School, he was a faculty in the School of Operations Research and Information Engineering at Cornell University. His research interests focus on revenue management, choice modeling, pricing, assortment optimization, and large-scale dynamic programming. From 2003 through 2004, he worked in the data mining and personalization group at Amazon.com. He received BA (1997) in Mathematics from University of California, Berkeley, and MS (1999) and PhD (2003) in Operations Research from Stanford University. He is a member of INFORMS.<\/p>\n"}],"event_agenda":false,"event_photo_gallery":false,"event_presentations":false,"event_custom_heading":[{"event_custom_title":"Abstract","event_custom_details":"
We develop a choice model that incorporates feedback from a population of customers making a purchase within the assortment of offered products. Each customer posts feedback on the product that she purchases within an offered product assortment. The feedback is obtained by applying a feedback function to the utility that the customer derives from her purchased product. When another customer is to make a purchase within the assortment, she uses the expected feedback from the customers with purchases to form a reference point. Therefore, the utility of a product has three components: intrinsic deterministic utility, Gumbel distributed idiosyncratic term and reference effect. The utility of a product is determined through a fixed point, as the utility depends on the reference effect, which, in turn, depends on the utilities of the products in the assortment. We give a closed-form expression for the choice probabilities for a broad class of feedback functions that can even have discontinuities. We develop an efficient algorithm to compute the revenue-maximizing assortment when the customers choose under our choice model. We study the pricing problem under our choice model, giving structural properties of the optimal prices, which ultimately yield an efficient algorithm for computing the revenue-maximizing prices. We also provide necessary and sufficient conditions for the convexity of the negative log-likelihood function. Our computational experiments demonstrate that our choice model can help predict the choices of the customers better than a number of benchmarks.<\/p>\n"}],"event_enquiry_details":{"event_enq_full_name":"","event_enq_department":"","event_enq_email":"","event_enq_telephone":"","event_enq_website":""}}]