ABSTRACT
Most hypervolume indicator based optimization algorithms like SIBEA [Zitzler et al. 2007], SMS-EMOA [Beume et al. 2007], or MO-CMA-ES [Igel et al. 2007] remove the solution with the smallest loss with respect to the dominated hypervolume from the population. This is usually iterated λ times until the size of the population no longer exceeds a fixed size μ. We show that there are populations such that the contributing hypervolume of the λ solutions chosen by this greedy selection scheme can be much smaller than the contributing hypervolume of an optimal set of λ solutions. Selecting the optimal λ-set implies calculating (μ+λ over μ)conventional hypervolume contributions, which is considered computationally too expensive. We present the first hypervolume algorithm which calculates directly the contribution of every set of λ solutions. This gives an additive term of (μ+λ over μ)in the runtime of the calculation instead of a multiplicative factor of binomial(μ+λ over μ). Given a population of size n=μ+λ, our algorithm can calculate a set of λ≥ solutions with minimal d-dimensional hypervolume contribution in time O(nd/2 log + n λ) for d > 2. This improves all previously published algorithms by a factor of order nmin(λ,,d/2 for d >3. Therefore even if we remove the solutions one by one greedily as usual, we gain a speedup factor of n for all d > 3.
- J. Bader and E. Zitzler. HypE: An Algorithm for Fast Hypervolume-Based Many-Objective Optimization. TIK Report 286, Institut für Technische Informatik und Kommunikationsnetze, ETH Zürich, Nov. 2008.Google Scholar
- J. L. Bentley. Algorithms for Klee's rectangle problems, 1977. Department of Computer Science, Carnegie Mellon University, Unpublished notes.Google Scholar
- N. Beume, C. Fonseca, M. López-Ibáñez, L. Paquete, and J. Vahrenhold. On the complexity of computing the hypervolume indicator. Technical report CI-235/07, Technical University of Dortmund, December 2007.Google Scholar
- N. Beume, B. Naujoks, and M. Emmerich. SMS-EMOA: Multiobjective selection based on dominated hypervolume. European Journal of Operational Research, 181(3):1653--1669, 2007.Google ScholarCross Ref
- N. Beume and G. Rudolph. Faster S-metric calculation by considering dominated hypervolume as Klee's measure problem. In Proc. Second International Conference on Computational Intelligence (IASTED'06), pages 233--238, 2006.Google Scholar
- L. Bradstreet, L. Barone, and L. While. Maximising hypervolume for selection in multi-objective evolutionary algorithms. In Proc. IEEE Congress on Evolutionary Computation (CEC'06), pages 6208--6215, 2006.Google ScholarCross Ref
- K. Bringmann and T. Friedrich. Approximating the volume of unions and intersections of high-dimensional geometric objects. In Proc. 19th International Symposium on Algorithms and Computation (ISAAC'08), volume 5369 of Lecture Notes in Computer Science, pages 436--447. Springer, 2008. Google ScholarDigital Library
- K. Bringmann and T. Friedrich. Approximating the least hypervolume contributor: NP-hard in general, but fast in practice. In Proc. 5th International Conference on Evolutionary Multi-Criterion Optimization (EMO'09), 2009. To appear, available from http://arxiv.org/abs/0812.2636. Google ScholarDigital Library
- D. Brockhoff and E. Zitzler. Improving hypervolume-based multiobjective evolutionary algorithms by using objective reduction methods. In Proc. IEEE Congress on Evolutionary Computation (CEC'07), pages 2086--2093, 2007.Google ScholarCross Ref
- K. Deb, L. Thiele, M. Laumanns, and E. Zitzler. Scalable multi-objective optimization test problems. In Proc. IEEE Congress on Evolutionary Computation (CEC'02), pages 825--830, 2002.Google ScholarCross Ref
- M. Emmerich, N. Beume, and B. Naujoks. An EMO algorithm using the hypervolume measure as selection criterion. In Proc. Third International Conference on Evolutionary Multi-Criterion Optimization (EMO'05), pages 62--76, 2005. Google ScholarDigital Library
- C. Igel, N. Hansen, and S. Roth. Covariance matrix adaptation for multi-objective optimization. Evol. Comput., 15(1):1--28, 2007. Google ScholarDigital Library
- J. Knowles, D. Corne, and M. Fleischer. Bounded archiving using the Lebesgue measure. In Proc. IEEE Congress on Evolutionary Computation (CEC'03), volume 4, pages 2490--2497, Dec 2003.Google ScholarCross Ref
- J. D. Knowles. Local-Search and Hybrid Evolutionary Algorithms for Pareto Optimization. PhD thesis, Department of Computer Science, University of Reading, UK, 2002.Google Scholar
- B. Naujoks, N. Beume, and M. T. M. Emmerich. Multi-objective optimisation using S-metric selection: application to three-dimensional solution spaces. In Proc. IEEE Congress on Evolutionary Computation (CEC'05), pages 1282--1289, 2005.Google ScholarCross Ref
- M. H. Overmars and C.-K. Yap. New upper bounds in Klee's measure problem. SIAM J. Comput., 20(6):1034--1045, 1991. Announced at ph29th Annual Symposium on Foundations of Computer Science (FOCS'88). Google ScholarDigital Library
- R. L. While, L. Bradstreet, L. Barone, and P. Hingston. Heuristics for optimizing the calculation of hypervolume for multi-objective optimization problems. In Proc. IEEE Congress on Evolutionary Computation (CEC'05), pages 2225--2232, 2005.Google Scholar
- R. L. While, P. Hingston, L. Barone, and S. Huband. A faster algorithm for calculating hypervolume. IEEE Trans. Evolutionary Computation, 10(1):29--38, 2006. Google ScholarDigital Library
- X. Zhou, N. Mao, W. Li, and C. Sun. A fast algorithm for computing the contribution of a point to the hypervolume. In Proc. Third International Conference on Natural Computation (ICNC'07), volume 4, pages 415--420, 2007. Google ScholarDigital Library
- E. Zitzler. Hypervolume metric calculation, 2001. Computer Engineering and Networks Laboratory (TIK), ETH Zurich, Switzerland, ftp://ftp.tik.ee.ethz.ch/pub/people/zitzler/hypervol.c.Google Scholar
- E. Zitzler, D. Brockhoff, and L. Thiele. The hypervolume indicator revisited: On the design of Pareto-compliant indicators via weighted integration. In Proc. Fourth International Conference on Evolutionary Multi-Criterion Optimization (EMO'07), volume 4403 of Lecture Notes in Computer Science, pages 862--876. Springer, 2007. Google ScholarDigital Library
- E. Zitzler and S. Künzli. Indicator-based selection in multiobjective search. In Proc. 8th International Conference Parallel Problem Solving from Nature (PPSN VIII), volume 3242 of Lecture Notes in Computer Science, pages 832--842. Springer, 2004.Google ScholarCross Ref
- E. Zitzler and L. Thiele. Multiobjective evolutionary algorithms: a comparative case study and the strength Pareto approach. IEEE Trans. Evolutionary Computation, 3(4):257--271, 1999. Announced at ph5th International Conference Parallel Problem Solving from Nature (PPSN V). Google ScholarDigital Library
- E. Zitzler, L. Thiele, M. Laumanns, C. M. Fonseca, and V. G. da Fonseca. Performance assessment of multiobjective optimizers: an analysis and review. IEEE Trans. Evolutionary Computation, 7(2):117--132, 2003. Google ScholarDigital Library
Index Terms
- Don't be greedy when calculating hypervolume contributions
Recommendations
Reference point specification in hypervolume calculation for fair comparison and efficient search
GECCO '17: Proceedings of the Genetic and Evolutionary Computation ConferenceHypervolume has been frequently used as a performance indicator for comparing evolutionary multiobjective optimization (EMO) algorithms. Hypervolume has been also used in indicator-based algorithms. Whereas a reference point is needed for hypervolume ...
Indicator-based evolutionary algorithm with hypervolume approximation by achievement scalarizing functions
GECCO '10: Proceedings of the 12th annual conference on Genetic and evolutionary computationPareto dominance-based algorithms have been the main stream in the field of evolutionary multiobjective optimization (EMO) for the last two decades. It is, however, well-known that Pareto-dominance-based algorithms do not always work well on many-...
A faster algorithm for calculating hypervolume
We present an algorithm for calculating hypervolume exactly, the Hypervolume by Slicing Objectives (HSO) algorithm, that is faster than any that has previously been published. HSO processes objectives instead of points, an idea that has been considered ...
Comments