skip to main content

Title: Four-Dimensional Golden Search

The Golden search technique is a method to search a multiple-dimension space to find the minimum. It basically subdivides the possible ranges of parameters until it brackets, to within an arbitrarily small distance, the minimum. It has the advantages that (1) the function to be minimized can be non-linear, (2) it does not require derivatives of the function, (3) the convergence criterion does not depend on the magnitude of the function. Thus, if the function is a goodness of fit parameter such as chi-square, the convergence does not depend on the noise being correctly estimated or the function correctly following the chi-square statistic. And, (4) the convergence criterion does not depend on the shape of the function. Thus, long shallow surfaces can be searched without the problem of premature convergence. As with many methods, the Golden search technique can be confused by surfaces with multiple minima.
Authors:
 [1]
  1. Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
Publication Date:
OSTI Identifier:
1171678
Report Number(s):
LA-UR--15-21424
DOE Contract Number:
AC52-06NA25396
Resource Type:
Technical Report
Research Org:
Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
Sponsoring Org:
USDOE
Country of Publication:
United States
Language:
English
Subject:
99 GENERAL AND MISCELLANEOUS