Hello, i am having a problem, i am not a pro in numerical computation so any help is welcome to solve my situation here. i have this Z(v) Function: http://dl.dropbox.com/u/6380744/img/zfunction.png
and i need to compute U(v): U(v) = (1+v)* d[ log(Z) ] / dv, in the range v [-1.0, 0.0] for small problems, we can compute U symbolically with Mathematica, and the result is: http://dl.dropbox.com/u/6380744/img/ufunction.png However, when i compute U with gsl_spline using gsl_interp_cspline, and gsl_spline_eval_deriv i get different values near v=-1.0 http://dl.dropbox.com/u/6380744/img/uNumeric.png i know that the problem comes because log(Z) diverges at v=-1.0, and numerically this is a problem. however when multiplied by (1+v), analitically the result is like the first U i showed solved in mathematica. A lot of further work depends on U, and when we deal with bigger problems, the numerical way will be our only solution. is there a way to solve this problem with the GSL capabilities?? any help is welcome, really! _______________________________________________ Help-gsl mailing list [email protected] http://lists.gnu.org/mailman/listinfo/help-gsl
