Generalized Response Surface Methodology: A New Metaheuristic

Research output: Working paperDiscussion paperOther research output

276 Downloads (Pure)

Abstract

Generalized Response Surface Methodology (GRSM) is a novel general-purpose metaheuristic based on Box and Wilson.s Response Surface Methodology (RSM).Both GRSM and RSM estimate local gradients to search for the optimal solution.These gradients use local first-order polynomials.GRSM, however, uses these gradients to estimate a better search direction than the steepest ascent direction used by RSM.Moreover, GRSM allows multiple responses, selecting one response as goal and the other responses as constrained variables.Finally, these estimated gradients may be used to test whether the estimated solution is indeed optimal.The focus of this paper is optimization of simulated systems.
Original languageEnglish
Place of PublicationTilburg
PublisherOperations research
Number of pages20
Volume2006-77
Publication statusPublished - 2006

Publication series

NameCentER Discussion Paper
Volume2006-77

Keywords

  • experimental design
  • multivariate regression analysis
  • least squares
  • Karush-Kuhn-Tucker conditions
  • bootstrap

Fingerprint Dive into the research topics of 'Generalized Response Surface Methodology: A New Metaheuristic'. Together they form a unique fingerprint.

Cite this