||August 5, 2009 10:15
Does any one know a good algorithm for auto-adjusting a relaxation factor? I have an algorithm that is having some problems reaching the solution at low residuals. So far, I have a small routine that adjusts the value of the relaxation coefficient based on the shakingness of the residuals for a variable (in this case gas temperature). I was wondering if someone knows a good paper or another reference for a better way to auto-adjust a relaxation factor.