search

hidmic / ltitop

Linear Time Invariant Topology Optimization Toolkit
GNU Lesser General Public License v3.0
0 stars 0 forks source link

Improve WCPG estimation technique #4

Open hidmic opened 3 years ago

hidmic commented 3 years ago

Worst case peak gain in ltitop is currently estimated by brute force:

https://github.com/hidmic/ltitop/blob/d75f165117d108ba5496b29c9920de04fa282c3d/src/ltitop/models/analysis.py#L45-L56

Accuracy guarantees are thus quite loose, and the resulting performance is awful. It calls for a better approach.

hidmic commented 3 years ago

I've attempted to use fixif.WCPG for this, but performance isn't great either and the requirement on the system matrix A to be diagonalizable is a bit limiting.