Learning optimization models in the presence of unknown relations

S. Verwer, Y. Zhang, Q.C. Ye

Research output: Contribution to journalArticleAcademic

Abstract

In a sequential auction with multiple bidding agents, it is highly challenging to determine the ordering of the items to sell in order to maximize the revenue due to the fact that the autonomy and private information of the agents heavily influence the outcome of the auction. The main contribution of this paper is two-fold. First, we demonstrate how to apply machine learning techniques to solve the optimal ordering problem in sequential auctions. We learn regression models from historical auctions, which are subsequently used to predict the expected value of orderings for new auctions. Given the learned models, we propose two types of optimization methods: a black-box best-first search approach, and a novel white-box approach that maps learned models to integer linear programs (ILP) which can then be solved by any ILP-solver. Although the studied auction design problem is hard, our proposed optimization methods obtain good orderings with high revenues. Our second main contribution is the insight that the internal structure of regression models can be efficiently evaluated inside an ILP solver for optimization purposes. To this end, we provide efficient encodings of regression trees and linear regression models as ILP constraints. This new way of using learned models for optimization is promising. As the experimental results show, it significantly outperforms the black-box best-first search in nearly all settings.
Original languageUndefined
Article number1401.1061v2 [cs.AI]
Number of pages37
JournalarXiv
Issue number1401.1061v2
Publication statusPublished - 6 Jan 2014
Externally publishedYes

Cite this

Verwer, S., Zhang, Y., & Ye, Q. C. (2014). Learning optimization models in the presence of unknown relations. arXiv, (1401.1061v2 ), [1401.1061v2 [cs.AI]].
Verwer, S. ; Zhang, Y. ; Ye, Q.C. / Learning optimization models in the presence of unknown relations. In: arXiv. 2014 ; No. 1401.1061v2 .
@article{0f21b01c0cda4dba9ecec1c7d37e6c8c,
title = "Learning optimization models in the presence of unknown relations",
abstract = "In a sequential auction with multiple bidding agents, it is highly challenging to determine the ordering of the items to sell in order to maximize the revenue due to the fact that the autonomy and private information of the agents heavily influence the outcome of the auction. The main contribution of this paper is two-fold. First, we demonstrate how to apply machine learning techniques to solve the optimal ordering problem in sequential auctions. We learn regression models from historical auctions, which are subsequently used to predict the expected value of orderings for new auctions. Given the learned models, we propose two types of optimization methods: a black-box best-first search approach, and a novel white-box approach that maps learned models to integer linear programs (ILP) which can then be solved by any ILP-solver. Although the studied auction design problem is hard, our proposed optimization methods obtain good orderings with high revenues. Our second main contribution is the insight that the internal structure of regression models can be efficiently evaluated inside an ILP solver for optimization purposes. To this end, we provide efficient encodings of regression trees and linear regression models as ILP constraints. This new way of using learned models for optimization is promising. As the experimental results show, it significantly outperforms the black-box best-first search in nearly all settings.",
keywords = "cs.AI, cs.GT, F.5.3; K.3; K.4",
author = "S. Verwer and Y. Zhang and Q.C. Ye",
year = "2014",
month = "1",
day = "6",
language = "Niet gedefinieerd",
journal = "arXiv",
publisher = "Cornell University Library",
number = "1401.1061v2",

}

Verwer, S, Zhang, Y & Ye, QC 2014, 'Learning optimization models in the presence of unknown relations', arXiv, no. 1401.1061v2 , 1401.1061v2 [cs.AI].

Learning optimization models in the presence of unknown relations. / Verwer, S.; Zhang, Y.; Ye, Q.C.

In: arXiv, No. 1401.1061v2 , 1401.1061v2 [cs.AI], 06.01.2014.

Research output: Contribution to journalArticleAcademic

TY - JOUR

T1 - Learning optimization models in the presence of unknown relations

AU - Verwer, S.

AU - Zhang, Y.

AU - Ye, Q.C.

PY - 2014/1/6

Y1 - 2014/1/6

N2 - In a sequential auction with multiple bidding agents, it is highly challenging to determine the ordering of the items to sell in order to maximize the revenue due to the fact that the autonomy and private information of the agents heavily influence the outcome of the auction. The main contribution of this paper is two-fold. First, we demonstrate how to apply machine learning techniques to solve the optimal ordering problem in sequential auctions. We learn regression models from historical auctions, which are subsequently used to predict the expected value of orderings for new auctions. Given the learned models, we propose two types of optimization methods: a black-box best-first search approach, and a novel white-box approach that maps learned models to integer linear programs (ILP) which can then be solved by any ILP-solver. Although the studied auction design problem is hard, our proposed optimization methods obtain good orderings with high revenues. Our second main contribution is the insight that the internal structure of regression models can be efficiently evaluated inside an ILP solver for optimization purposes. To this end, we provide efficient encodings of regression trees and linear regression models as ILP constraints. This new way of using learned models for optimization is promising. As the experimental results show, it significantly outperforms the black-box best-first search in nearly all settings.

AB - In a sequential auction with multiple bidding agents, it is highly challenging to determine the ordering of the items to sell in order to maximize the revenue due to the fact that the autonomy and private information of the agents heavily influence the outcome of the auction. The main contribution of this paper is two-fold. First, we demonstrate how to apply machine learning techniques to solve the optimal ordering problem in sequential auctions. We learn regression models from historical auctions, which are subsequently used to predict the expected value of orderings for new auctions. Given the learned models, we propose two types of optimization methods: a black-box best-first search approach, and a novel white-box approach that maps learned models to integer linear programs (ILP) which can then be solved by any ILP-solver. Although the studied auction design problem is hard, our proposed optimization methods obtain good orderings with high revenues. Our second main contribution is the insight that the internal structure of regression models can be efficiently evaluated inside an ILP solver for optimization purposes. To this end, we provide efficient encodings of regression trees and linear regression models as ILP constraints. This new way of using learned models for optimization is promising. As the experimental results show, it significantly outperforms the black-box best-first search in nearly all settings.

KW - cs.AI

KW - cs.GT

KW - F.5.3; K.3; K.4

M3 - Tijdschriftartikel

JO - arXiv

JF - arXiv

IS - 1401.1061v2

M1 - 1401.1061v2 [cs.AI]

ER -

Verwer S, Zhang Y, Ye QC. Learning optimization models in the presence of unknown relations. arXiv. 2014 Jan 6;(1401.1061v2 ). 1401.1061v2 [cs.AI].