Hi I posted earlier, been reading a bit since then. Still Im not sure of what method is best suited for the following problem... I use a CCD algorithm for finding a pose. At each joint I have to find the best attempt to rotate focus position to the wished locator position, considering the joints constraints. The joint is only rotational, but it can in worst case rotate about its own xyz axis, within limited intervals on these axis. I need a good method finding the "closest" possible position for the focus. Just calculating a desired rotation quaternion, convert it into euler and adjusting the values doesnt give this position. Searching through like this: goal = ..calc goal //unconstrained for X_MIN to X_MAX for Y_MIN to Y_MAX for Z_MIN to Z_MAX { focus = ... calc focus, based on current angles; distance = || goal - focus || if(distance < min_distance) { min_distance = distance; angles = (x,y,z); } } Its a bit slow for good results... How do u guys solve problems like this?....I saw something about Lagrange multipliers, but it was a bit too advanced for me I suppose, unless they are easier than they seem to be once explained properly. Or are there better ways? It has to be kinda fast, not supah-accurate and preferable easy to understand and implement.