The d-Level Nested Logit Model


Paat Rusmevichientong
USC Marshall School of Business

Wednesday, Oct.1, 2014
4:00 - 5:00 PM
GSB C102


Abstract:

We provide a new tree-based formulation for the d-level nested logit model, allowing for an arbitrary number of levels. We then consider the canonical revenue management problem of finding a revenue-maximizing assortment under this choice model. By exploiting the succinct descriptions of the selection probabilities and expected revenues under our formulation, we develop an efficient algorithm for computing an optimal assortment. For a d-level nested logit model with n products, the running time of the algorithm is O(d n log n). This is joint work with Guang Li and Huseyin Topaloglu.

---

Operations Research Colloquia: http://or.stanford.edu/oras_seminars.html