Hi, I am trying to understand how BFGS minimisation works, I have read through the chapter on minimisation in Numerical Recipes in C, although it's a bit confusing, I get the jist of the technique. But when I actually try and do something practical with it I hit a brick wall. The problem I am trying to solve is that of lobe fitting with Zonal Harmonics. Basically there is an error function which calculates how well a particular lobe fits a signal, and this is the function that is to be minimised. The function is dependant upon two variables, s* the lobe axis (A Scalar Quantity) and g* the lobe coefficients (A Vector Quantity). So the function is something like this E(g*, s*) = ... And i have two gradient functions (with respect to each of the variables) dE --- = ... dg* dE --- = ... ds* But I don't really see how 2D functions work with BFGS, is there something I am missing, or does anyone have any links that could help me. Many Thanks EDIT : Here are some links in case you find them useful : The paper I'm trying to understand Specifically the section on single lobe fitting using BFGS Numerical Recipes in C