Defines exact adaptivity gaps of e/(e-1) and 2 between policy classes for two-sided assortment optimization and provides 1/4-approximation for adaptive one-by-one policies plus 0.067-approximation for simultaneous policies under MNL, with extensions to constrained settings.
Let us denote by C t j the observed backlog of supplier j ∈ S and by Mt the number of matches in the t-th run
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.OC 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Two-sided Assortment Optimization: Adaptivity Gaps and Approximation Algorithms
Defines exact adaptivity gaps of e/(e-1) and 2 between policy classes for two-sided assortment optimization and provides 1/4-approximation for adaptive one-by-one policies plus 0.067-approximation for simultaneous policies under MNL, with extensions to constrained settings.