MathGroup Archive 2002

[Date Index] [Thread Index] [Author Index]

Search the Archive

Marquardt & Levenberg Method Theoretical Origins

  • To: mathgroup at smc.vnet.net
  • Subject: [mg38404] Marquardt & Levenberg Method Theoretical Origins
  • From: tasande at MIT.EDU
  • Date: Fri, 13 Dec 2002 04:19:02 -0500 (EST)
  • Sender: owner-wri-mathgroup at wolfram.com

Dear Mathematicons:

In reference to the recent inquiries concerning the theoretical basis
of the infamous Marquardt/Levenberg Method, I enjoin you to first read
the seminal paper:

Marquardt, D. "An Algorithm for Least-Squares Estimation of Nonlinear
Parameters." SIAM J. Appl. Math. 11, 431-441, 1963. 

Marduart reformulates the original Levenberg method and introduces his
eponymous "compromise".  The paper is clear and cogent.  It will help
you immensely in your subsequent selection of a modern implementation
of the algorithm.

You may also wish to consult the pure Levenberg Method in the ancient paper:

Levenberg, K. "A Method for the Solution of Certain Problems in Least Squares."
Quart. Appl. Math. 2, 164-168, 1944.

Sincerely,

Theodore Sande
MIT Department of Physics




  • Prev by Date: Re: What is -- Fractional integro-differentiation
  • Next by Date: Re: too many {{}}'s for listplot
  • Previous by thread: Strange Output Involving Times, Power
  • Next by thread: looking for a "test[ 1, ALL ]