重みの適応を行うスカラー化勾配法による多目的関数最適化

カテゴリ: 部門大会

論文No: TC13-5

グループ名: 【C】平成22年電気学会電子・情報・システム部門大会講演論文集

発行日: 2010/09/02

タイトル(英語): Multiobjective Function Optimization Using a Scalarized Descent Method with a Weight Adaptation

著者名: 濱田 直希(東京工業大学),永田 裕一(東京工業大学),小林 重信(東京工業大学),小野 功(東京工業大学)

著者名(英語): Naoki Hamada(Tokyo Institute of Technology),Yuichi Nagata(Tokyo Institute of Technology),Shigenobu Kobayashi(Tokyo Institute of Technology),Isao Ono(Tokyo Institute of Technology)

キーワード: 多目的関数最適化|多数目的問題|スカラー化|勾配法|重みの適応|multiobjective function optimization|many-objective problem|scalarization|descent method|weight adaptation

要約(日本語): The multi-starting descent method is a promising approach to unimodal multiobjective function optimization problems because of its precision of obtained solutions. The scalarized descent method scalarizes objective functions to a single function with a weight vector and uses the gradient of the scalarized function. Its convergent point depends on a weight vector. Unfortunately, in conventional multi-starting strategies, it is difficult to choose appropriate weight vectors for obtaining widely and evenly distributed solutions. In order to remedy the problems, we propose a multi-starting scalarized descent method named AWA that employs an adaptive scheme of weight vectors. We show the effectiveness of the proposed method through some experiments.

PDFファイルサイズ: 7,733 Kバイト