Submodular Order Functions and Assortment Optimization

by   Rajan Udwani, et al.

We define a new class of set functions that in addition to being monotone and subadditive, also admit a very limited form of submodularity defined over a permutation of the ground set. We refer to this permutation as a submodular order. This class of functions includes monotone submodular functions as a sub-family. To understand the importance of this structure in optimization problems we consider the problem of maximizing function value under various types of constraints. To demonstrate the modeling power of submodular order functions we show applications in two different settings. First, we apply our results to the extensively studied problem of assortment optimization. While the objectives in assortment optimization are known to be non-submodular (and non-monotone) even for simple choice models, we show that they are compatible with the notion of submodular order. Consequently, we obtain new and in some cases the first constant factor guarantee for constrained assortment optimization in fundamental choice models. As a second application of submodular order functions, we show an intriguing connection to the maximization of monotone submodular functions in the streaming model. We recover some best known guarantees for this problem as a corollary of our results.



There are no comments yet.


page 1

page 2

page 3

page 4


Greed is Not Always Good: On Submodular Maximization over Independence Systems

In this work, we consider the maximization of submodular functions const...

On Submodular Prophet Inequalities and Correlation Gap

Prophet inequalities and secretary problems have been extensively studie...

Approximate Submodular Functions and Performance Guarantees

We consider the problem of maximizing non-negative non-decreasing set fu...

DOPE: Distributed Optimization for Pairwise Energies

We formulate an Alternating Direction Method of Mul-tipliers (ADMM) that...

On the Approximation Relationship between Optimizing Ratio of Submodular (RS) and Difference of Submodular (DS) Functions

We demonstrate that from an algorithm guaranteeing an approximation fact...

Fully-Dynamic Submodular Cover with Bounded Recourse

In submodular covering problems, we are given a monotone, nonnegative su...

Parallel Quasi-concave set optimization: A new frontier that scales without needing submodularity

Classes of set functions along with a choice of ground set are a bedrock...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.