Re: conjugate gradient stuck at inflection point

On Aug 22, 6:11 pm, Scott Seidman <namdiestt...@xxxxxxxxxxxxxx> wrote:
Palle Uldenborg <palle.uldenb...@xxxxxxxxx> wrote innews:1187776523.289408.6290@xxxxxxxxxxxxxxxxxxxxxxxxxxx:

Next, take a search for "simulated annealing" methods. Imagine that you're
minimization is just a marble sliding down a surface. The idea is to give
the marble a bit of a push to help it over the hills, then taper off the
extra energy.

Hi Scott

Thank you for the reply, but perhaps I was unclear.

I am trying to minimize a function named U(x). In other words I am
looking for a local minimum if U(x), but the algorithm converges to a
local minimum of the function W(x)=|grad U(x)|^2

Palle

.

Quantcast