Tight bounds for the approximation ratio of the hypervolume indicator

13Citations
Citations of this article
13Readers
Mendeley users who have this article in their library.
Get full text

Abstract

The hypervolume indicator is widely used to guide the search and to evaluate the performance of evolutionary multi-objective optimization algorithms. It measures the volume of the dominated portion of the objective space which is considered to give a good approximation of the Pareto front. There is surprisingly little theoretically known about the quality of this approximation. We examine the multiplicative approximation ratio achieved by two-dimensional sets maximizing the hypervolume indicator and prove that it deviates significantly from the optimal approximation ratio. This provable gap is even exponential in the ratio between the largest and the smallest value of the front. We also examine the additive approximation ratio of the hypervolume indicator and prove that it achieves the optimal additive approximation ratio apart from a small factor ≤ n/(n - 2), where n is the size of the population. Hence the hypervolume indicator can be used to achieve a very good additive but not a good multiplicative approximation of a Pareto front. © 2010 Springer-Verlag.

Cite

CITATION STYLE

APA

Bringmann, K., & Friedrich, T. (2010). Tight bounds for the approximation ratio of the hypervolume indicator. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6238 LNCS, pp. 607–616). https://doi.org/10.1007/978-3-642-15844-5_61

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free