Publications · 2010 · Conference paper
Alternative Fitness Assignment Methods for Many-Objective Optimization Problems
In Artifical Evolution, 9th International Conference, Evolution Artificielle, EA 2009, 2010
Abstract
Pareto dominance (PD) has been the most commonly adopted relation to compare solutions in the multiobjective optimization context. Multiobjective evolutionary algorithms (MOEAs) based on PD have been successfully used in order to optimize bi-objective and three-objective problems. However, it has been shown that Pareto dominance loses its effectiveness as the number of objectives increases and thus, the convergence behavior of approaches based on this concept decreases. This paper tackles the MOEAs’ scalability problem that arises as we increase the number of objective functions. In this paper, we perform a comparative study of some of the state-of-the-art fitness assignment methods available for multiobjective optimization in order to analyze their ability to guide the search process in high-dimensional objective spaces.
BibTeX
@inproceedings{Fabre10,
author = {Mario Garza-Fabre and Gregorio Toscano Pulido and Carlos A. {Coello Coello}},
title = {{Alternative Fitness Assignment Methods for Many-Objective Optimization Problems}},
booktitle = {Artifical Evolution, 9th International Conference, Evolution Artificielle, EA 2009},
pages = {146--157},
publisher = {Springer. Lecture Notes in Computer Science, Vol. 5975},
address = {Strasbourg, France},
year = {2010},
doi = {10.1007/978-3-642-14156-0_13},
url = {https://doi.org/10.1007/978-3-642-14156-0_13},
isbn = {978-3-642-14155-3},
}