國立虎尾科技大學 |

Optimal Mechanism Design in Sequential Decision Making Processes.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Optimal Mechanism Design in Sequential Decision Making Processes./
Author:	Lyu, Boxiang.
Description:	1 online resource (261 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-09, Section: B.
Contained By:	Dissertations Abstracts International85-09B.
Subject:	Computer science. -
Online resource:	click for full text (PQDT)
ISBN:	9798381943412

Optimal Mechanism Design in Sequential Decision Making Processes.
Lyu, Boxiang.

Optimal Mechanism Design in Sequential Decision Making Processes. - 1 online resource (261 pages)

Source: Dissertations Abstracts International, Volume: 85-09, Section: B.

Thesis (Ph.D.)--The University of Chicago, 2024.

Includes bibliographical references

For the dissertation, we propose studying efficiently learning the optimal dynamic mechanism when the agents' valuations can be characterized by a Markov Decision Process (MDP). In Chapter 1, we provide a high-level motivation for our problem setting, discussing the motivations for combining theoretical Reinforcement Learning (RL) with dynamic mechanism design. In subsequent chapters, we provide three representative problems at the intersection of these fields. These chapters vary in terms of the difficulty of designing the optimal mechanism, the generality of function approximation, and the RL setup considered, providing a high-level overview of recent advances along this interdisciplinary research direction.In particular, in Chapter 2, we show how the revenue maximizing, incentive compatible, and ex-post individually rational mechanism can be learned computationally and sample efficiently, when the single buyer's type distribution is governed by a tabular MDP. In Chapter 3, we provide an online learning algorithm that learns the optimal second-price auction with reserve prices with $\ilde{O}(\\sqrt{T})$ regret, even if the participating buyers behave strategically. In Chapter 4, we show that the welfare-maximizing, ex-ante incentive compatible, and ex-ante individually rational mechanism can be obtained under general function approximation setting using only a pre-collected data set, with no additional interactions with the environment.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798381943412Subjects--Topical Terms:

573171
Computer science.
Subjects--Index Terms:

Algorithmic game theoryIndex Terms--Genre/Form:

554714
Electronic books.

Optimal Mechanism Design in Sequential Decision Making Processes.
LDR:02794ntm a22003737 4500 001 1148347
005 20240924101900.5
006 m o d
007 cr bn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798381943412
035 $a (MiAaPQ)AAI30990395
035 $a AAI30990395
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Lyu, Boxiang. $3 1474293
245 1 0 $a Optimal Mechanism Design in Sequential Decision Making Processes.
264 0 $c 2024
300 $a 1 online resource (261 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-09, Section: B.
500 $a Advisor: Kolar, Mladen.
502 $a Thesis (Ph.D.)--The University of Chicago, 2024.
504 $a Includes bibliographical references
520 $a For the dissertation, we propose studying efficiently learning the optimal dynamic mechanism when the agents' valuations can be characterized by a Markov Decision Process (MDP). In Chapter 1, we provide a high-level motivation for our problem setting, discussing the motivations for combining theoretical Reinforcement Learning (RL) with dynamic mechanism design. In subsequent chapters, we provide three representative problems at the intersection of these fields. These chapters vary in terms of the difficulty of designing the optimal mechanism, the generality of function approximation, and the RL setup considered, providing a high-level overview of recent advances along this interdisciplinary research direction.In particular, in Chapter 2, we show how the revenue maximizing, incentive compatible, and ex-post individually rational mechanism can be learned computationally and sample efficiently, when the single buyer's type distribution is governed by a tabular MDP. In Chapter 3, we provide an online learning algorithm that learns the optimal second-price auction with reserve prices with $\ilde{O}(\\sqrt{T})$ regret, even if the participating buyers behave strategically. In Chapter 4, we show that the welfare-maximizing, ex-ante incentive compatible, and ex-ante individually rational mechanism can be obtained under general function approximation setting using only a pre-collected data set, with no additional interactions with the environment.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Computer science. $3 573171
650 4 $a Statistics. $3 556824
653 $a Algorithmic game theory
653 $a Dynamic mechanism design
653 $a Theoretical reinforcement learning
655 7 $a Electronic books. $2 local $3 554714
690 $a 0984
690 $a 0511
690 $a 0463
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a The University of Chicago. $b Business. $3 1179073
773 0 $t Dissertations Abstracts International $g 85-09B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30990395 $z click for full text (PQDT)