Ellipsoids for contextual dynamic pricing

Maxime C. Cohen, Ilan Lobel, Renato Paes Leme

Abstract

We study the dynamic pricing problem faced by a firm selling differentiated products. At each period, the firm receives a new product, which is described by a vector of features. The firm needs to choose prices, but it does not know a priori the market value of the different features. We first consider an algorithm that we call PolytopePricing, but prove that it incurs worst-case regret that scales exponentially in the dimensionality of the feature space. We then consider a closely related algorithm, EllipsoidPricing, and show it incurs low regret with regards to both the time horizon and the dimensionality of the feature space. For more details, we refer the reader to our full paper.
BibTeX
@article{cohen2017ellipsoids,
  author = {Maxime C. Cohen and Ilan Lobel and Renato Paes Leme},
  title = {{Ellipsoids for contextual dynamic pricing}},
  journal = {Sigecom Exch},
  year = {2017},
  doi = {10.1145/3055589.3055595},
}

Loading related content...

Figure