Presentation 1998/5/21
Optimization Problems from Genetic Algorithms' Viewpoints
Akiko Aizawa,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper gives a brief introduction to some of the theoretical works in genetic algorithms and discuss their relationship and the future possibilities. We specifically focus on the problem of solution space characterization of fixed length bit strings and examine the mathematical implication of these three representative characterization methods: Walsh analysis, Epsitasis variance, crossover correlation. All these methods assume that the solution space is modeled as a linear combination of sub-strings and that the performance of the search can be represented by the fitting of the model. However, it is also known that such criteria dose not always match with the actual behavior of genetic algorithms. As future research directions, two approaches are introduced, one based on a fitness landscape model, and the other related to existing experimental design or global random search theory.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Genetic Algorithms / Fitness Landscape / Walsh Analysis / Epistasis / Fitness Correlation / Design of Experiments
Paper #
Date of Issue

Conference Information
Committee AI
Conference Date 1998/5/21(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Artificial Intelligence and Knowledge-Based Processing (AI)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Optimization Problems from Genetic Algorithms' Viewpoints
Sub Title (in English)
Keyword(1) Genetic Algorithms
Keyword(2) Fitness Landscape
Keyword(3) Walsh Analysis
Keyword(4) Epistasis
Keyword(5) Fitness Correlation
Keyword(6) Design of Experiments
1st Author's Name Akiko Aizawa
1st Author's Affiliation National Center for Science Information Systems()
Date 1998/5/21
Paper #
Volume (vol) vol.98
Number (no) 58
Page pp.pp.-
#Pages 8
Date of Issue