PDF4PRO ⚡AMP

Modern search engine that looking for books and documents around the web

Example: marketing

METHOD OF QUADRATIC INTERPOLATION

METHOD OF QUADRATIC INTERPOLATIONKELLER methods are a common approach to the more generalarea of line search for optimization. In the case of QUADRATIC inter-polation, the function s critical value is bracketed, and a quadraticinterpolant is fitted to the arc contained in the interval. Then, theinterpolant is minimized, and the new interval is determined based onthe relation of the minimizer to the original endpoints of the more formally, letx* maximize (or minimize)f(x). Ifx* isnot easily found through analytic methods, then it is significantly eas-ier to bracket the interval over which this critical point occurs. Letq(x) denote the QUADRATIC interpolant off(x). Minimizing a quadraticfunction is trivial, and so the critical point ofqis easily obtained. Wethen form a new bracketing interval by throwing away the worst point, which for our purposes would be the point that is the largest orsmallest, depending on whether we want to approximate a maximumor minimum.

This remainder term will often be denoted R n+1. Theorem 3.3. Let f: R !R be 3 times continuously di erentiable. Let x* be such that f0(x) = 0 and f00(x) 6= 0 . Then the sequence fx ng generated by (2.7) converges to x with order 1+ p 5 2. Proof. We rst want to prove that (2.7) does indeed converge to our minimizer, x.

Loading..

Tags:

  Theorem, Remainder

Information

Domain:

Source:

Link to this page:

Please notify us if you found a problem with this document:

Spam in document Broken preview Other abuse

Transcription of METHOD OF QUADRATIC INTERPOLATION

Related search queries