Selection mechanisms based on the maximin fitness function to solve multi-objective optimization problems

详细信息    查看全文

• 作者：Adriana Menchaca-Mendez ; ^{adriana.menchacamendez@gmail.com" class="auth_mail" title="E-mail the corresponding author} ; Carlos A. Coello Coello ^{ccoello@cs.cinvestav.mx" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Multi-objective evolutionary algorithms ; Selection operators ; Maximin fitness function
• 刊名：Information Sciences
• 出版年：2016
• 出版时间：1 March 2016
• 年：2016
• 卷：332
• 期：Complete
• 页码：131-152
• 全文大小：5668 K

文摘

In this paper, we study three selection mechanisms based on the maximin fitness function and we propose another one. These selection mechanisms give rise to the following MOEAs: “MC-MOEA”, “MD-MOEA”, “MH-MOEA” and “MAH-MOEA”. We validated them using standard test functions taken from the specialized literature, having from three up to ten objective functions. We compare these four MOEAs among them and also with respect to MOEA/D (which is based on decomposition), and to SMS-EMOA (which is based on the hypervolume indicator). Our preliminary results indicate that “MD-MOEA” and “MAH-MOEA” are promising alternatives for solving MOPs with either low or high dimensionality.