Services & Resources / Wolfram Forums / MathGroup Archive
-----

MathGroup Archive 2012

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

Search the Archive

Obtaining the last evaluation parameters when NMaximize does not converge

  • To: mathgroup at smc.vnet.net
  • Subject: [mg128399] Obtaining the last evaluation parameters when NMaximize does not converge
  • From: felipe.benguria at gmail.com
  • Date: Sat, 13 Oct 2012 01:02:39 -0400 (EDT)
  • Delivered-to: l-mathgroup@mail-archive0.wolfram.com
  • Delivered-to: l-mathgroup@wolfram.com
  • Delivered-to: mathgroup-newout@smc.vnet.net
  • Delivered-to: mathgroup-newsend@smc.vnet.net

Dear all,

 I am maximizing a function with ~ 100 parameters using NMaximize.  I cannot   find a maximum.  It would be helpful if once the function stops with an error (or I stop it when it is not making progress) I could obtain the last set of parameters that were used to evaluate the function being maximized.

I understand that reap / sow shows the whole series of points where the function was evaluated , once the maximum is found, so I have not been able to use it.

In this case, given how slow my maximization goes, it would be good to find a method that does not slow down the code too much.

I appreciate your help

Felipe



  • Prev by Date: Re: Re: Grouping graphics that are transformed together
  • Next by Date: Re: D under Sum
  • Previous by thread: Resetting Clock[]
  • Next by thread: sum of coins article in mathematica journal