cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Community Tip - Want the oppurtunity to discuss enhancements to PTC products? Join a working group! X

Optimization methods/algorithms

ptc-5277933
1-Newbie

Optimization methods/algorithms

What methods or algorithms are available in Creo when conducting an optimization study?

If descriptions of these methods are not commonly known a brief explanation would be greatly appreciated.

5 REPLIES 5

To be clear on what I am asking I attached the screenshot below. The preferences window is found through optimization/feasibility >> options >> preferences.

I am asking how the GDP and MDS methods work. I have had no success searching for these abreaveations if these are a common method the full name would be helpfull. If these are propritary or modified methods it would be helpfull to know what they are similar to or baised on.

optimization_prefrences.jpg

I think it depends on the context of your analysis....mechanism or structural, etc. According to Creo 2.0 help topics in the context of mechanism (I'm assuming that looks like what you are doing):

  • Gradient Projection (GDP). Use the standard algorithm to optimize the model using the current model conditions as the starting point.
  • Multi-Objective Design Study (MDS). Use the multi-objective design studies algorithm to determine the optimum starting point for the optimization. You can specify the number of starting points to compute in the Max Iterations field. This method has a higher chance of finding the overall optimum design within the design parameters and dimensions.

It won't matter what type of measures you're using, mechanisms, Simulate, etc. In general, the GDP method is more likely to find a local optimum, due to the fact that it starts at current values and conducts one search. The MDS method breaks-up your ranges for each variable and sweeps a more thorough range that is much more likely to find your global optimum because it performs multiple searches starting in many locations. Of course, the MDS method does take more time.

Thanks, for the replies. Your help is greatly appreciated!

I have Creo 2.0 M100.

Its seems that they have changed the labels of the algorithms or what types are availble. Above they talk about GDP and MDS. It seems like MDS is a global optimization which is what i need. I think i am having problems finding the overall best solution. I just keep getting local solutions based on the starting point and limits.

However, MDS is no longer available or i dont know where to find it. The options i see for algorithm are GDP and SQP. The descriptions ive found for these only talk about whether or the not the solution dimensions are feasible. But it doesnt say anything about the global solution or starting points.

Optimization Algorithm

Use Optimization Algorithm in the Design Study Options dialog box to specify an algorithm to the Creo Simulate engine other than the default algorithm for an optimization design study. When making this decision, you should understand the advantages and disadvantages of each algorithm.

Automatic—If you select Automatic option, Creo Simulate begins the run using the default SQP algorithm. However, if Creo Simulate encounters an invalid model during an optimization and fails to recover the model it attempts to resolve the problem by automatically switching from SQP to GDP for the remainder of the run.
Sequential Quadratic Programming (SQP)—Use this option to find the optimum design faster than GDP does. The disadvantage of SQP is that it does not guarantee that your design satisfies your limits at the end of each iteration. It only guarantees that the optimum design satisfies your limits. This means that if SQP ever fails to find an optimum design, there may be no intermediate designs available that are improvements over the initial design.
Gradient Projection (GDP)—Use this option only if the speed is not an issue and you want to ensure the availability of interim designs. GDP tends to produce a series of intermediate designs that satisfy your limits while getting closer to the goal.

Is something like MDS still available? Do you have any suggestions for finding global solutions to multi variable optimizations in Creo Simulate?

Announcements
Business Continuity with Creo: Learn more about it here.

Top Tags