Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
1993
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*December
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 1993

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

Search the Archive

Re: alternative to FindMinimum

  • To: mathgroup at yoda.physics.unc.edu
  • Subject: Re: alternative to FindMinimum
  • From: keiper
  • Date: Wed, 5 May 93 09:20:51 CDT

	> i have a "function" that i would like to minimize but
	> FindMinimum does not work on it because there are no derivatives.

I don't know what is meant by "there are no derivatives".  If it really
is a nowhere differentiable function I suspect that it will be very difficult
to minimize numerically regardless of the algorithm used.  If it means that
D[ ] does not produce a symbolic derivative then "differentiability" is
irrelevant: FindMinimum does not require derivatives:

In[21]:= FindMinimum[Abs[x]^2, {x, 1, 2}]

                    -28                  -14
Out[21]= {1.51868 10   , {x -> 1.23235 10   }}


(Note that it does assume that the function is smooth.  If the function is
not smooth convergence will be much slower.  Squaring Abs[x] is one way to
make it smooth.)

Jerry B. Keiper
keiper at wri.com
Wolfram Research, Inc.





  • Prev by Date: alternative to FindMinimum
  • Next by Date: Stepwise Regression?
  • Previous by thread: alternative to FindMinimum
  • Next by thread: Stepwise Regression?