東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

State representation for agent-based...

Mao, Tao.

FindBook

Google Book

Amazon

博客來

State representation for agent-based reinforcement learning.

紀錄類型:	書目-語言資料,印刷品 : Monograph/item
正題名/作者:	State representation for agent-based reinforcement learning./
作者:	Mao, Tao.
面頁冊數:	201 p.
附註:	Source: Dissertation Abstracts International, Volume: 75-03(E), Section: B.
Contained By:	Dissertation Abstracts International75-03B(E).
標題:	Engineering, Computer. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3601713
ISBN:	9781303533136

State representation for agent-based reinforcement learning.
Mao, Tao.

State representation for agent-based reinforcement learning. - 201 p.

Source: Dissertation Abstracts International, Volume: 75-03(E), Section: B.

Thesis (Ph.D.)--Dartmouth College, 2013.

Agent-based policy learning in complex and uncertain environments is challenged by escalating computational complexity with the size of the task space (action choices and environmental states) as well as the number of agents. Nonetheless, there is ample evidence in the natural world that high functioning social mammals learn to solve complex problems with ease, both individually and cooperatively. Neuroscientific research shows that brain structures embed computational functions that include state abstraction, hierarchical state representation, and action space chaining.

ISBN: 9781303533136Subjects--Topical Terms:

1669061
Engineering, Computer.

State representation for agent-based reinforcement learning.
LDR:03074nam a2200313 4500 001 1966838
005 20141112075118.5
008 150210s2013 ||||||||||||||||| ||eng d
020 $a 9781303533136
035 $a (MiAaPQ)AAI3601713
035 $a AAI3601713
040 $a MiAaPQ $c MiAaPQ
100 1 $a Mao, Tao. $3 2103722
245 1 0 $a State representation for agent-based reinforcement learning.
300 $a 201 p.
500 $a Source: Dissertation Abstracts International, Volume: 75-03(E), Section: B.
500 $a Adviser: Laura E. Ray.
502 $a Thesis (Ph.D.)--Dartmouth College, 2013.
520 $a Agent-based policy learning in complex and uncertain environments is challenged by escalating computational complexity with the size of the task space (action choices and environmental states) as well as the number of agents. Nonetheless, there is ample evidence in the natural world that high functioning social mammals learn to solve complex problems with ease, both individually and cooperatively. Neuroscientific research shows that brain structures embed computational functions that include state abstraction, hierarchical state representation, and action space chaining.
520 $a Drawing from these themes, this thesis develops two important concepts for reducing the state complexity: state abstraction (mapping an external state vector to an internal state representation of lower dimension) and hierarchical state representation (representing a state space by a hierarchy of subspaces thereby reducing the number of states used to represent the environment). The Q-Tree algorithm, which provides simultaneous state abstraction, hierarchical representation, and policy learning in a single algorithm, is introduced, requiring no prior knowledge of candidate boundaries for separating the state space into subspaces when solving agent-based reinforcement learning problems. The Q-Tree algorithm automatically constructs linear separations on multiple dimensions and has a computational complexity linear in the dimension of the state vector and independent of the potential number of state boundaries, which is superior to existing methods of constructing hierarchical state representations for reinforcement learning.
520 $a These methodologies are applied to single- and multi-agent task domains such as foraging, patrolling, herding and path planning so as to demonstrate performance in terms of convergence rate, computational efficiency, memory requirements and generalizability of learned policies. The Q-Tree algorithm is also employed to model incubation and restructuring in insight problem solving, accounting for three-stage learning curves observed in rhesus monkeys in a reverse-reward contingency task and generalizing learned knowledge to new but similar situations.
590 $a School code: 0059.
650 4 $a Engineering, Computer. $3 1669061
650 4 $a Artificial Intelligence. $3 769149
650 4 $a Engineering, Electronics and Electrical. $3 626636
690 $a 0464
690 $a 0800
690 $a 0544
710 2 $a Dartmouth College. $b Engineering. $3 1057396
773 0 $t Dissertation Abstracts International $g 75-03B(E).
790 $a 0059
791 $a Ph.D.
792 $a 2013
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3601713