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

Near optimal vehicle dispatching: Co...

Petroff, Thomas M.

FindBook

Google Book

Amazon

博客來

Near optimal vehicle dispatching: Combining linear programming and reinforcement learning.

紀錄類型:	書目-語言資料,印刷品 : Monograph/item
正題名/作者:	Near optimal vehicle dispatching: Combining linear programming and reinforcement learning./
作者:	Petroff, Thomas M.
面頁冊數:	87 p.
附註:	Adviser: Nicholas S. Flann.
Contained By:	Masters Abstracts International46-01.
標題:	Artificial Intelligence. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=1445070
ISBN:	9780549115847

Near optimal vehicle dispatching: Combining linear programming and reinforcement learning.
Petroff, Thomas M.

Near optimal vehicle dispatching: Combining linear programming and reinforcement learning. - 87 p.

Adviser: Nicholas S. Flann.

Thesis (M.S.)--Utah State University, 2007.

Continuous vehicle dispatching is an NP-complete problem, thus requiring local methods to solve the problem. This paper presents a multi-stage approach to solving the continuous vehicle dispatching problem. In the first stage, linear programming is used to create an abstract schedule. In the second stage, a Monte Carlo reinforcement learning algorithm is used to operationalize the schedule, achieving near optimal results. The reinforcement learning algorithm uses the abstract schedule as its state representation. Tests are performed to compare this method against another multi-stage approach using linear programming and a load balancing local search method. The reinforcement learning approach is shown to perform better than the load balancing local search method.

ISBN: 9780549115847Subjects--Topical Terms:

769149
Artificial Intelligence.

Near optimal vehicle dispatching: Combining linear programming and reinforcement learning.
LDR:01665nam 2200289 a 45 001 959033
005 20110704
008 110704s2007 ||||||||||||||||| ||eng d
020 $a 9780549115847
035 $a (UMI)AAI1445070
035 $a AAI1445070
040 $a UMI $c UMI
100 1 $a Petroff, Thomas M. $3 1282501
245 1 0 $a Near optimal vehicle dispatching: Combining linear programming and reinforcement learning.
300 $a 87 p.
500 $a Adviser: Nicholas S. Flann.
500 $a Source: Masters Abstracts International, Volume: 46-01, page: 0391.
502 $a Thesis (M.S.)--Utah State University, 2007.
520 $a Continuous vehicle dispatching is an NP-complete problem, thus requiring local methods to solve the problem. This paper presents a multi-stage approach to solving the continuous vehicle dispatching problem. In the first stage, linear programming is used to create an abstract schedule. In the second stage, a Monte Carlo reinforcement learning algorithm is used to operationalize the schedule, achieving near optimal results. The reinforcement learning algorithm uses the abstract schedule as its state representation. Tests are performed to compare this method against another multi-stage approach using linear programming and a load balancing local search method. The reinforcement learning approach is shown to perform better than the load balancing local search method.
590 $a School code: 0241.
650 4 $a Artificial Intelligence. $3 769149
650 4 $a Computer Science. $3 626642
650 4 $a Engineering, Mining. $3 1035560
690 $a 0551
690 $a 0800
690 $a 0984
710 2 $a Utah State University. $3 960049
773 0 $t Masters Abstracts International $g 46-01.
790 $a 0241
790 1 0 $a Flann, Nicholas S., $e advisor
791 $a M.S.
792 $a 2007
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=1445070