Return
A new multiobjective evolutionary algorithm
DOI:10.1016/S0377-2217(01)00190-4.png)
Abstract
En 中文
The Pareto-based approaches have shown some success in designing multiobjective evolutionary algorithms (MEAs). Their methods of fitness assignment are mainly from the information of dominated and nondominated individuals. On the top of the hierarchy of MEAs, the strength Pareto evolutionary algorithm (SPEA) has been elaborately designed with this principle in mind. In this paper, we propose a (mu+lambda) multiobjective evolutionary algorithm ((mu+lambda) MEA), which discards the dominated individuals in each generation. The comparisons of the experimental results demonstrate that the (mu+lambda) MEA outperforms SPEA on five benchmark functions with less computational efforts. (C) 2002 Elsevier Science B.V. All rights reserved.
Journal
IF:
6
Papers:
2.2W
Citations:
8.1W
Organization
No organization information available

