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

Learnability analysis of real-coded ...

Zhang, Jian.

Linked to FindBook

Google Book

Amazon

博客來

Learnability analysis of real-coded evolutionary algorithms.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Learnability analysis of real-coded evolutionary algorithms./
Author:	Zhang, Jian.
Description:	107 p.
Notes:	Source: Dissertation Abstracts International, Volume: 65-07, Section: B, page: 3556.
Contained By:	Dissertation Abstracts International65-07B.
Subject:	Computer Science. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3140684
ISBN:	0496878247

Learnability analysis of real-coded evolutionary algorithms.
Zhang, Jian.

Learnability analysis of real-coded evolutionary algorithms. - 107 p.

Source: Dissertation Abstracts International, Volume: 65-07, Section: B, page: 3556.

Thesis (Ph.D.)--Tulane University, 2004.

Evolutionary Algorithms is a school of algorithms inspired by biological evolutionary progress. Darwin's Principle of Evolution, Over many generations, random variation and nature selection shape the behaviors of individuals and species to fit the demands of their surroundings, is the intrinsic concept of Evolutionary Algorithms design. Despite of their practical success, more theoretical questions for better understanding on their mechanism are needed to be answered. For example, the convergence rate of EAs, the population sizing and genetic operator selection, are questions remaining open.

ISBN: 0496878247Subjects--Topical Terms:

626642
Computer Science.

Learnability analysis of real-coded evolutionary algorithms.
LDR:02307nmm 2200277 4500 001 1846718
005 20051103093540.5
008 130614s2004 eng d
020 $a 0496878247
035 $a (UnM)AAI3140684
035 $a AAI3140684
040 $a UnM $c UnM
100 1 $a Zhang, Jian. $3 1057605
245 1 0 $a Learnability analysis of real-coded evolutionary algorithms.
300 $a 107 p.
500 $a Source: Dissertation Abstracts International, Volume: 65-07, Section: B, page: 3556.
500 $a Chair: Bill P. Buckles.
502 $a Thesis (Ph.D.)--Tulane University, 2004.
520 $a Evolutionary Algorithms is a school of algorithms inspired by biological evolutionary progress. Darwin's Principle of Evolution, Over many generations, random variation and nature selection shape the behaviors of individuals and species to fit the demands of their surroundings, is the intrinsic concept of Evolutionary Algorithms design. Despite of their practical success, more theoretical questions for better understanding on their mechanism are needed to be answered. For example, the convergence rate of EAs, the population sizing and genetic operator selection, are questions remaining open.
520 $a Population size is usually an empirical parameter when adjusting Evolutionary Algorithms. In this investigation, we study the population sizing problem from two new aspects: fitness landscapes' ruggedness and the Probably Approximately Correct (PAC) learning. The population size is therefore also referred to as sample complexity using the language of statistical learning. We characterize the population sizing as a learning process of the search space so that the Evolutionary Algorithms has high probability to globally optimize the fitness function. We define granularity, a ruggedness measure for fitness functions. We also derive a sampling theorem based on the epsilon-cover and the PAC learning theory, which theoretically determines a population size towards effective convergence for global optimization and multimodal optimization tasks.
590 $a School code: 0235.
650 4 $a Computer Science. $3 626642
690 $a 0984
710 2 0 $a Tulane University. $3 1019475
773 0 $t Dissertation Abstracts International $g 65-07B.
790 1 0 $a Buckles, Bill P., $e advisor
790 $a 0235
791 $a Ph.D.
792 $a 2004
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3140684