A Method For Approximating Univariate Convex Functions Using Only Function Value Evaluations

A.Y.D. Siem, D. den Hertog, A.L. Hoffmann

Research output: Working paperDiscussion paperOther research output

475 Downloads (Pure)

Abstract

In this paper, piecewise linear upper and lower bounds for univariate convex functions are derived that are only based on function value information. These upper and lower bounds can be used to approximate univariate convex functions. Furthermore, new Sandwich algo- rithms are proposed, that iteratively add new input data points in a systematic way, until a desired accuracy of the approximation is obtained. We show that our new algorithms that use only function-value evaluations converge quadratically under certain conditions on the derivatives. Under other conditions, linear convergence can be shown. Some numeri- cal examples, including a Strategic investment model, that illustrate the usefulness of the algorithm, are given.
Original languageEnglish
Place of PublicationTilburg
PublisherOperations research
Number of pages26
Volume2007-67
Publication statusPublished - 2007

Publication series

NameCentER Discussion Paper
Volume2007-67

Keywords

  • approximation
  • convexity
  • meta-model
  • Sandwich algorithm

Fingerprint

Dive into the research topics of 'A Method For Approximating Univariate Convex Functions Using Only Function Value Evaluations'. Together they form a unique fingerprint.

Cite this