[Help-gsl] multimin: gsl_multimin_fdfminimizer_iterate( ) set variables and value of target function to "NaN"

[kevin]

Hi,

I'm using the conjugate_fr algorithm in an optimization problem, 

However, after the program executes the function "
gsl_multimin_fdfminimizer_iterate (s)" , 

all the variables "x" (i.e.gsl_vector_get (s->x, k)) and the value of target
function (i.e. s->f ) are set to NaN, and they remain NaN in the following
iterations.

Does anyone know what is the potential problem ? 

My target function includes digamma fuction, which is taylor approximation
of first derivative of the log gamma function. I guess maybe it cause the
problem.

If anyone who like to help me need my primary code, please contact me!

 

Thanks all,

 

Kevin