search

artofscience / SAOR

Sequential Approximate Optimization Repository
GNU General Public License v3.0
5 stars 1 forks source link

Dynamic compliance problem definition #82

Open Giannis1993 opened 2 years ago

Giannis1993 commented 2 years ago

This problem needs some polishing as it doesn't give black/white design. Not sure what it needs, but I remember @aatmdelissen said it had something to do with the problem definition.

artofscience commented 2 years ago

@aatmdelissen what to do with the dynamic topopt problem. Do we want to polish it or remove it?

aatmdelissen commented 2 years ago

The dynamic compliance on its own does not converge nicely, but we could add it as a constraint for the eigenfrequency example.

artofscience commented 2 years ago

The dynamic compliance on its own does not converge nicely, but we could add it as a constraint for the eigenfrequency example.

Isnt it a good example then to show a very hard problem?

aatmdelissen commented 2 years ago

What I meant was that the optimum in dynamic compliance by itself is not uniquely defined; there are many local optima with a 0 dynamic compliance at one frequency Silva, et al. (2019). That's why you need an additional objective function such as compliance, or maybe eigenfrequencies.

artofscience commented 2 years ago

I see, then adding the dynamic compliance as a constraint would be nice? (or other way around)