<会議発表論文>
Comparison of Mapping Methods to Visualize the EC Landscape
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 | |
関連DOI | |
概要 | We compare four mapping methods -self-organizing map (SOM), Sammon's non-linear mapping (NLM), topology preserving mapping of sample sets (TOPAS), and VISOR algorithm -to visualize the landscape of ev...olutionary computation (EC) and accelerate the convergence of EC and interactive EC (IEC). Three experiments are conducted using five benchmark functions and 28 subjects. We compare the computational complexity, the capacity for human visualization, and the effect on convergence for each experiment. These experiments showed that SOM demonstrated the best performance, and the VISOR performed well when CPU time was critical. The other mapping methods, NLM and TOPAS, were far from practical.続きを見る |
目次 | 1 Introduction 2 Mapping Algorithms 3 Visualized GA as an Experimental System 4 Experimental Comparison of Mapping Methods 5 Conclusion |
本文ファイル
ファイル | ファイルタイプ | サイズ | 閲覧回数 | 説明 |
---|---|---|---|---|
IntConf070 | 1.83 MB | 114 |
詳細
NCID | |
---|---|
レコードID | |
関連ISBN | |
登録日 | 2021.08.27 |
更新日 | 2021.08.27 |