[ADMB Users] Are parameters scaled for optimization.

dave fournier davef at otter-rsch.com
Wed Jan 8 07:59:39 PST 2014


In all cases the function minimizer just sees an unbound variable x. It 
is what haapens to x in between the
function minimizer and the model being parameterized  that determines 
the boundedness and or scaling.
for example

          x ------> a+(b-a)*sin(2*pi*x)

is bounded between a and b from the models point of view

         x ------->x/s

has been scaled by s from the function minimizers point of view

         x ------> a+(b-a)*sin(2*pi*x/s)

is bounded from the models point of view and scaled by s from the
function minimizers point of view.



More information about the Users mailing list