Change Soft convergence criterion

Altair Forum User
Altair Forum User
Altair Employee
edited October 2020 in Community Q&A

Hello,

 

when i start an optimization with optistruct the calculation is finished after 1 Iteration. In the Output-file i found following Information:

 

Soft convergence criterion satisfied;
the design did not change during the last Iteration.

 

What does this mean?

 

I already did a lot of successful shape-optimizations but in this case i don't really know why the shape didn't change.

I changed the default OBJTOL value from the opti control card but the problem still occurs.

The most interesting thing is:

When i change the inital value in the shape panel to 0.5 I don't have any problems and the optimization finds the best shape. When I start with an inital value of 0.0 then i get the problem with the soft convergence criterion.    

 

Thank you

Greets,

 

Fatih Uysal

Answers

  • Rahul Rajan_21763
    Rahul Rajan_21763 New Altair Community Member
    edited November 2020

    Soft convergence is achieved when there is little or no change in the design variables for two consecutive iterations. It is not necessary to evaluate the objective (or constraints) for the final design point, as the model is unchanged from the previous iteration. Therefore, soft convergence requires one less iteration than regular.

    Attached document should help you in understanding the process in detail

    Unable to find an attachment - read this blog

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited May 2016

    Thank you Rahul for the Information.

    Since Optistruct is a gradient based optimizer i guess optistruct stopped optimization after finding a local optimum.

     

    Using a DGLOBAL-Card helped me to find the global optimum.

     

    Greets,

     

    Fatih Uysal   

  • Rahul Rajan_21763
    Rahul Rajan_21763 New Altair Community Member
    edited May 2016

    Thanks for the update.