Creo Simulate > Reference Links > Strategy: After You Run an Optimization Study
  
Strategy: After You Run an Optimization Study
You can follow up optimization studies with standard and local sensitivity studies in the following ways:
If you ran an optimization study with a high analysis convergence percentage to save time, the values for the design variables are probably valid even though the goal and limit values may not be accurate.
You can then lower the convergence value in the analysis you included in that study and run a standard study with your design variables set to their optimization values. This study will give you more accurate results for the goal and other quantities.
You can also run a new optimization study after lowering the convergence values. Use the final position of your design variables in the old study as their initial position in the new study. The new optimization will produce more accurate results and could further refine the optimized shape of your model.
If any design variables in an optimization need to meet a standard size or other manufacturing requirement, you can set those design variables to the standard size closest to the optimized value and run a standard study. You can then check the results of the standard study to see if other quantities of interest are still close to their optimized value.
If an optimization study ends with a message in the report file that says changes in the goal quantity were insignificant relative to its initial value, you can use local sensitivity to check the goal quantity. Set the start position of each design variable to the optimized position.
After the sensitivity study completes, graph the goal quantity against each design variable, and check to see whether the slope of each graph is close to zero. A slope that is not near zero indicates that the optimization study may not have reached the optimum goal value. In this case, you may want to redefine the optimization study and run it again.
If you run the optimization again, use the design variable values from the final optimum model of the last optimization study as the starting point in the new study.
If an optimization study ends with insignificant changes to your design variables, Creo Simulate may have encountered a local optimum value for your goal measure that caused it not to explore the design space more fully.
If this happens, run the study again after setting the design variables to initial values far enough from the original settings to encourage Creo Simulate to examine more promising design variable positions.
Return to Strategy: Optimizing a Model.