<会議発表論文>
Learning Control Strategies for High Performance Genetic Algorithms

作成者
本文言語
出版者
発行日
収録物名
開始ページ
終了ページ
会議情報
出版タイプ
アクセス権
概要 In this paper, we propose a method to learn high performance strategies for controlling genetic algorithms. In our proposal, control strategies are represented by fuzzy systems that dynamically contro...l population sizing, crossover rates, and mutation rates. The control strategies are acquired and optimized according to online and of Hine measures using a genetic algorithm technique. We compare control strategies obtained using our methods with optimized static genetic algorithms and show performance Improvements. In some experiments, these strategies use a combination of high crossover rates, fluctuating population size, and exponentially decreasing mutation rates to realize high online and of Hine performance.続きを見る
目次 1 Introduction
2 The Dynamic Parametric GA
3 Results
4 Conclusions and Further Research

本文ファイル

pdf IntConf027 pdf 1.72 MB 131  

詳細

レコードID
登録日 2021.08.24
更新日 2021.08.24