ABSTRACT
Model- and simulation-designers are often interested not in the optimum output of their system, but in understanding how the output is sensitive to different parameters. This can require an inefficient sweep of a multidimensional parameter space, with many samples tested in regions of the space where the output is essentially all the same, or a sparse sweep which misses crucial "interesting" regions where the output is strongly sensitive. In this paper we introduce a novel population-oriented approach to adaptive parameter sweeping which focuses its samples on these sensitive areas. The method is easy to implement and model-free, and does not require any previous knowledge about the space. In a weakened form the method can operate in non-metric spaces such as the space of genetic program trees. We demonstrate the method on three test problems, showing that it identifies regions of the space where the slope of the output is highest, and concentrates samples on those regions.
- N. M. Amato and Y. Wu. A randomized roadmap method for path and manipulation planning. In IEEE International Conference on Robotics and Automation (ICRA), 1996.Google ScholarCross Ref
- S. Baluja. Population-based incremental learning: A method for integrating genetic search based function optimization and competitive learning. Technical Report CMU-CS-94-163, Carnegy Mellon University, Pittsburgh, PA, 1994. Google ScholarDigital Library
- S. Ghosh and C. R. Rao, editors. Design and Analysis of Experiments, volume 13 of Handbook of Statistics. Elsevier Science, 1996.Google Scholar
- G. R. Harik, F. G. Lobo, and D. E. Goldberg. The compact genetic algorithm. In Proceedings of the IEEE Conference on Evolutionary Computation, 1998.Google ScholarCross Ref
- D. Hsu, T. Jiang, J. Reif, and Z. Sun. The bridge test for sampling narrow passages with probabilistic roadmap planners. In IEEE International Conference on Robotics and Automation (ICRA), 2003.Google ScholarCross Ref
- P. Larrañaga, R. Etxeberria, J. A. Lozano, and J. M. Peña. Optimization in continuous domains by learning and simulation of gaussian networks. In Optimization By Building and Using Probabilistic, pages 201--204, Las Vegas, Nevada, USA, 8 July 2000.Google Scholar
- P. Larrañaga, J. A. Lozano, and E. Bengoetxea. Estimation of distribution algorithms based on multivariate normal and gaussian networks. Technical Report KZZA-IK-1-01, Department of Computer Science and Artificial Intelligence, University of the Basque Country, Spain, 2001.Google Scholar
- M. McKay, R. Beckman, and W. Conover. A comparison of three methods for selecting values of input variables in the analysis of output from a compuer code. Technometrics, 21(2), 1979.Google Scholar
- D. C. Montgomery. Design and Analysis of Experiments. John Wiley, New-York, 2005. Google ScholarDigital Library
- H. Mühlenbein. The equation for response to selection and its use for prediction. Evolutionary Computation, 5(3), 1997. Google ScholarDigital Library
- M. Pelikan, D. E. Goldberg, and E. Cantó-Paz. BOA: The Bayesian optimization algorithm. In W. Banzhaf, J. Daida, A. E. Eiben, M. H. Garzon, V. Honavar, M. Jakiela, and R. E. Smith, editors, Proceedings of the Genetic and Evolutionary Computation Conference GECCO-99, volume I, pages 525--532, Orlando, FL, 1999. Morgan Kaufmann Publishers, San Fransisco, CA.Google Scholar
- S. Rudlof and M. Koppen. Stochastic hill climbing by vectors of normal distributions. In Proceedings of the First Online Workshop on Soft Computing, WSC1, 1996.Google Scholar
- G. G. Wang. Adaptive response surface method using inherited latin hypercube design points. ASME Transactions, Journal of Mechanical Design, 125(2):210--220, 2003.Google ScholarCross Ref
- G. G. Wang, Z. Dong, and P. Aitchison. Adaptive response surface method a global optimization scheme for computation-intensive design problems. Journal of Engineering Optimization, 33(6):707--734, 2001.Google ScholarCross Ref
- Y. S. Yeun, B. J. Kim, Y. S. Yang, and W. S. Ruy. Polynomial genetic programming for response surface modeling part 2: adaptive approximate models with probabilistic optimization problems. Structural and Multidisciplinary Optimization, 29(1):35--49, Jan. 2005.Google ScholarCross Ref
- Y. S. Yeun, Y. S. Yang, W. S. Ruy, and B. J. Kim. Polynomial genetic programming for response surface modeling part 1: a methodology. Structural and Multidisciplinary Optimization, 29(1):19--34, Jan. 2005.Google ScholarCross Ref
Index Terms
- Finding interesting things
Recommendations
On finding a widest empty 1-corner corridor
A 1-corner corridor through a set S of points is an open subset of CH(S) containing no points from S and bounded by a pair of parallel polygonal lines each of which contains two segments. Given a set of n points in the plane, we consider the problem of ...
Finding the Maximum Empty Axis-Parallel Rectangular Annulus
Combinatorial Image AnalysisAbstractAn annulus is basically a ring-shaped region between two concentric disks on the same plane. However, it can be defined on any other geometrical shapes, for example, a rectangular annulus is defined as the area between two rectangles with one ...
On finding widest empty curved corridors
An α-siphon of width w is the locus of points in the plane that are at the same distance w from a 1-corner polygonal chain C such that α is the interior angle of C. Given a set P of n points in the plane and a fixed angle α, we want to compute the ...
Comments