An adaptive data structure for evolutionary multi-objective algorithms with unbounded archives

Author

Yuen, Joseph ; Gao, Sophia ; Wagner, Markus ; Neumann, Frank

Author_Institution

Sch. of Comput. Sci., Univ. of Adelaide, Adelaide, SA, Australia

fYear

2012

fDate

10-15 June 2012

Firstpage

1

Lastpage

8

Abstract

Archives have been widely used in evolutionary multi-objective optimization in order to store the optimal points found so far during the optimization process. Usually the size of an archive is bounded which means that the number of points it can store is limited. This implies that knowledge about the set of non-dominated solutions that has been obtained during the optimization process gets lost. Working with unbounded archives allows to keep this knowledge which can be useful for the progress of an evolutionary multi-objective algorithm. In this paper, we propose an adaptive data structure for dealing with unbounded archives. This data structure allows to traverse the archive efficiently and can also be used for sampling solutions from the archive which can be used for reproduction.

Keywords

data structures; evolutionary computation; adaptive data structure; evolutionary multiobjective algorithms; evolutionary multiobjective optimization; optimization process; unbounded archives; Complexity theory; Data structures; Evolutionary computation; Optimization; Partitioning algorithms; Runtime; Vectors; Archive; Data Structures; Evolutionary Algorithm; Multi-Objective Optimization;

fLanguage

English

Publisher

ieee

Conference_Titel

Evolutionary Computation (CEC), 2012 IEEE Congress on

Conference_Location

Brisbane, QLD

Print_ISBN

978-1-4673-1510-4

Electronic_ISBN

978-1-4673-1508-1

Type

conf

DOI

10.1109/CEC.2012.6256468

Filename

6256468