Volume: 18, Issue: 2(2009)
pp. 355-362 DOI: 10.1142/S0218213009000160
|
|
Abstract |
Full Text (PDF, 321KB)
|
References
|
 |
| Title: |
UNIFORMLY DISTRIBUTED PARETO FRONTS THROUGH THE MAXIMIN SELECTION ALGORITHM |
| Author(s): |
SERGIO I. VALDEZ PEÑA
Department of Computer Science, Centre for Research in Mathematics (CIMAT), A.P. 402, Guanajuato, Gto., MéxicoSALVADOR BOTELLO RIONDA
Department of Computer Science, Centre for Research in Mathematics (CIMAT), A.P. 402, Guanajuato, Gto., MéxicoARTURO HERNÁNDEZ AGUIRRE
Department of Computer Science, Centre for Research in Mathematics (CIMAT), A.P. 402, Guanajuato, Gto., México
|
| History: |
Received 29 November 2007
Accepted 24 September 2008
|
| Abstract: |
This paper introduces the MaxiMin selection algorithm, a deterministic procedure to achieve maximal spread and almost perfectly distributed Pareto fronts. MaxiMin is successful because the measure of uniformity is performed on the set being constructed, not on the source set as most multi-objective algorithms do. For comparison purposes we present results delivered by several MOEAs with and without the MaxiMin selection. Performance metrics and graphical results show that MaxiMin improves the distribution and spread of Pareto fronts, with no negative effects on the convergence performance of the multi-objective algorithm. |
| Keywords: |
Pareto front diversity; multi-objective; uniform spreading; MOEAs
|
|
|
