作成者 |
|
|
|
本文言語 |
|
出版者 |
|
|
|
発行日 |
|
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 |
To make interactive GA practical, we have tried to reduce psychological/physical burden of human operators by improving human interface and GA convergence. This paper proposes a method that combines G...A and modeling human evaluation to make GA convergence better. In this proposed method, a quadratic function is fitted to a solution space in each generation, and the best individual obtained from the approximated function is replaced with worst individual obtained from real fitness function. We evaluated how the proposed method can make convergence better through simulations using five fitness functions of DeJong and drawing faces system with interactive GA. The results show that the proposed method can make GA convergence better in four functions of DeJong and drawing faces task and at least does not make the convergence poorer.続きを見る
|