#118: SOL200 – Why is Scaling Important in Optimization?
When performing an optimization analysis, it is not unusual to have many constraints and/or design variables that have large differences in magnitudes. If these values are used without scaling, some constraints and/or design variables may dominate. In some cases, the convergence rate may be slower, in other cases, the optimum design may not be reached at all. Normalizing the constraints and design variables also help the numerics of the problem.
Recent Comments