國立高雄大學圖資館 |

Language: English

Back

Double-Oracle Deep Reinforcement Learning for Handling Exponential Action Space in Sequential Stackelberg Security Games.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Double-Oracle Deep Reinforcement Learning for Handling Exponential Action Space in Sequential Stackelberg Security Games.
Author:	Tan, Czander.
Published:	Ann Arbor : ProQuest Dissertations & Theses, 2022
Description:	41 p.
Notes:	Source: Masters Abstracts International, Volume: 84-05.
Notes:	Advisor: Nguyen, Thanh H.
Contained By:	Masters Abstracts International84-05.
Subject:	Computer science.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29252657
ISBN:	9798352946459

Double-Oracle Deep Reinforcement Learning for Handling Exponential Action Space in Sequential Stackelberg Security Games.
Tan, Czander.

Double-Oracle Deep Reinforcement Learning for Handling Exponential Action Space in Sequential Stackelberg Security Games. - Ann Arbor : ProQuest Dissertations & Theses, 2022 - 41 p.

Source: Masters Abstracts International, Volume: 84-05.

Thesis (M.S.)--University of Oregon, 2022.

This item must not be sold to any third party vendors.

Standard Stackelberg Security Games (SSGs) assume attackers to be myopic players that select only a single target based on the defender's strategy. In this paper, we consider sequential SSGs, in which attackers launch multiple attacks sequentially. With a sequence of events, however, the defender's action space grows exponentially with the number of time steps, making the problem computationally intractable.To handle this issue, this paper presents the following contributions. First, we use the Double Oracle algorithm to iteratively derive player strategies. Second, we use Advantage Actor-Critic models to approximate best response policies for both players. Lastly, we represent the defender action space compactly with marginal probabilities instead of enumerating all possible actions.Overall, our experiments show that the Double Oracle algorithm not only allows us to search through defender strategies effectively and efficiently, but also provides optimal solutions that outperform other models at scalable settings.

ISBN: 9798352946459Subjects--Topical Terms:

199325
Computer science.
Subjects--Index Terms:

Double-oracle deep reinforcement learning

Double-Oracle Deep Reinforcement Learning for Handling Exponential Action Space in Sequential Stackelberg Security Games.
LDR:02179nmm a2200349 4500 001 636131
005 20230501063907.5
006 m o d
007 cr#unu||||||||
008 230724s2022 ||||||||||||||||| ||eng d
020 $a 9798352946459
035 $a (MiAaPQ)AAI29252657
035 $a AAI29252657
040 $a MiAaPQ $c MiAaPQ
100 1 $a Tan, Czander. $3 942470
245 1 0 $a Double-Oracle Deep Reinforcement Learning for Handling Exponential Action Space in Sequential Stackelberg Security Games.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2022
300 $a 41 p.
500 $a Source: Masters Abstracts International, Volume: 84-05.
500 $a Advisor: Nguyen, Thanh H.
502 $a Thesis (M.S.)--University of Oregon, 2022.
506 $a This item must not be sold to any third party vendors.
520 $a Standard Stackelberg Security Games (SSGs) assume attackers to be myopic players that select only a single target based on the defender's strategy. In this paper, we consider sequential SSGs, in which attackers launch multiple attacks sequentially. With a sequence of events, however, the defender's action space grows exponentially with the number of time steps, making the problem computationally intractable.To handle this issue, this paper presents the following contributions. First, we use the Double Oracle algorithm to iteratively derive player strategies. Second, we use Advantage Actor-Critic models to approximate best response policies for both players. Lastly, we represent the defender action space compactly with marginal probabilities instead of enumerating all possible actions.Overall, our experiments show that the Double Oracle algorithm not only allows us to search through defender strategies effectively and efficiently, but also provides optimal solutions that outperform other models at scalable settings.
590 $a School code: 0171.
650 4 $a Computer science. $3 199325
653 $a Double-oracle deep reinforcement learning
653 $a Exponential action space
690 $a 0984
690 $a 0800
710 2 $a University of Oregon. $b Department of Computer and Information Science. $3 942471
773 0 $t Masters Abstracts International $g 84-05.
790 $a 0171
791 $a M.S.
792 $a 2022
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29252657