EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
3354. Data-Driven Submodular Set-function Optimization: Theory and Applications in Assortment Planning and Recommender Systems
Invited abstract in session TD-3: Data science meets strongly NP-Hard CO , stream Data Science Meets Optimization.
Tuesday, 14:30-16:00Room: 1005 (building: 202)
Authors (first author is the speaker)
| 1. | Jigar Patel
|
| Management Science, University of Miami |
Abstract
Recommender systems and informed assortment planning facilitate the display of goods or services tailored to individual customer needs/characteristics, promote the visibility of assorted products, and revenue growth. We address the problem of optimal assortment and display of online search results for goods or services. The objective is to maximize the platform's revenue and customer engagement by leveraging the menu of displayed search results to users, subject to catering to their individual search criteria or characteristics. Our analysis is based on a detailed data set from a leading online platform. Assortment planning involves optimization over a utility set-function. For example, a supermodular utility function is related to complementary goods, and a submodular utility function is related to substitutes. By analyzing our data, we identify properties of our utility function and implement the proper optimization algorithm tailored to individual users.
Recent advances in the discrete optimization field has enabled us to provide optimal or near-optimal algorithms in polynomial time to tackle the difficult problem of submodular function optimization
Keywords
- Analytics and Data Science
- Inventory
- Stochastic Optimization
Status: accepted
Back to the list of papers