東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Using artificial intelligence techni...

Ko, Ilsang.

Linked to FindBook

Google Book

Amazon

博客來

Using artificial intelligence techniques and learning to solve multi-level knapsack problems.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Using artificial intelligence techniques and learning to solve multi-level knapsack problems./
Author:	Ko, Ilsang.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 1993,
Description:	211 p.
Notes:	Source: Dissertations Abstracts International, Volume: 55-01, Section: A.
Contained By:	Dissertations Abstracts International55-01A.
Subject:	Computer science. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=9320442
ISBN:	9798208664285

Using artificial intelligence techniques and learning to solve multi-level knapsack problems.
Ko, Ilsang.

Using artificial intelligence techniques and learning to solve multi-level knapsack problems. - Ann Arbor : ProQuest Dissertations & Theses, 1993 - 211 p.

Source: Dissertations Abstracts International, Volume: 55-01, Section: A.

Thesis (Ph.D.)--University of Colorado at Boulder, 1993.

In solving multi-level knapsack problems, conventional heuristic approaches often assume a shortsighted plan within a static decision environment to find a near optimal solution. These conventional approaches are inflexible, and lack the ability to adapt to different problem structures. This research approaches the problem from a totally different viewpoint, and a new method that utilizes artificial intelligence (AI) techniques is designed and implemented. The term "AI technique" implies that one performs intelligent actions based on memories of historic data, knowledge bases, and learning. These actions are developed not only by observing the attributes of the optimal solution, the solution space, and its corresponding path to the optimal solution, but also by applying human intelligence, experience, and intuition with respect to the search strategies. By accomplishing memory-based learning from previous history, the intensive learning approach makes decisions in a dynamic way. The approach intensifies, or diversifies the search process appropriately in time and space. In order to create a good neighborhood structure, this learning approach uses two powerful choice rules that emphasize the impact of candidate variables on the current solution with respect to their profit contribution. A side effect of so-called "pseudo moves," similar to "aspirations," supports these choice rules during the evaluation process. For the purpose of visiting as many relevant points as possible, strategic oscillation between feasible and infeasible solutions around the boundary is applied for intensification. To avoid redundant moves, short-term (tabu-lists), intermediate-term (cycle detection), and long-term (recording frequency and significant solutions for diversification) memories are used. In addition, the concepts of "diversification within intensification" and "intensification within diversification" also enhance the intensive learning approach by dynamically creating a effective rhythm of the search process. From a macroperspective, all the intelligent actions are systematically applied during the oscillation between the "information-acquiring process" and the "information-utilizing process." Test results show that among the 45 generated problems (these problems pose significant or insurmountable challenges to exact methods) the approach produces the optimal solutions in 39 cases.

ISBN: 9798208664285Subjects--Topical Terms:

523869
Computer science.

Using artificial intelligence techniques and learning to solve multi-level knapsack problems.
LDR:03506nmm a2200349 4500 001 2399668
005 20240916075430.5
006 m o d
007 cr#unu||||||||
008 251215s1993 ||||||||||||||||| ||eng d
020 $a 9798208664285
035 $a (MiAaPQ)AAI9320442
035 $a AAI9320442
040 $a MiAaPQ $c MiAaPQ
100 1 $a Ko, Ilsang. $3 677184
245 1 0 $a Using artificial intelligence techniques and learning to solve multi-level knapsack problems.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 1993
300 $a 211 p.
500 $a Source: Dissertations Abstracts International, Volume: 55-01, Section: A.
500 $a Publisher info.: Dissertation/Thesis.
500 $a Advisor: Kelly, James;Glover, Fred.
502 $a Thesis (Ph.D.)--University of Colorado at Boulder, 1993.
520 $a In solving multi-level knapsack problems, conventional heuristic approaches often assume a shortsighted plan within a static decision environment to find a near optimal solution. These conventional approaches are inflexible, and lack the ability to adapt to different problem structures. This research approaches the problem from a totally different viewpoint, and a new method that utilizes artificial intelligence (AI) techniques is designed and implemented. The term "AI technique" implies that one performs intelligent actions based on memories of historic data, knowledge bases, and learning. These actions are developed not only by observing the attributes of the optimal solution, the solution space, and its corresponding path to the optimal solution, but also by applying human intelligence, experience, and intuition with respect to the search strategies. By accomplishing memory-based learning from previous history, the intensive learning approach makes decisions in a dynamic way. The approach intensifies, or diversifies the search process appropriately in time and space. In order to create a good neighborhood structure, this learning approach uses two powerful choice rules that emphasize the impact of candidate variables on the current solution with respect to their profit contribution. A side effect of so-called "pseudo moves," similar to "aspirations," supports these choice rules during the evaluation process. For the purpose of visiting as many relevant points as possible, strategic oscillation between feasible and infeasible solutions around the boundary is applied for intensification. To avoid redundant moves, short-term (tabu-lists), intermediate-term (cycle detection), and long-term (recording frequency and significant solutions for diversification) memories are used. In addition, the concepts of "diversification within intensification" and "intensification within diversification" also enhance the intensive learning approach by dynamically creating a effective rhythm of the search process. From a macroperspective, all the intelligent actions are systematically applied during the oscillation between the "information-acquiring process" and the "information-utilizing process." Test results show that among the 45 generated problems (these problems pose significant or insurmountable challenges to exact methods) the approach produces the optimal solutions in 39 cases.
590 $a School code: 0051.
650 4 $a Computer science. $3 523869
690 $a 0454
690 $a 0796
690 $a 0984
690 $a 0800
710 2 $a University of Colorado at Boulder. $3 1019435
773 0 $t Dissertations Abstracts International $g 55-01A.
790 $a 0051
791 $a Ph.D.
792 $a 1993
793 $a English
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=9320442