search

NanoComp / photonics-opt-testbed

testbed problems for photonics optimization
MIT License
27 stars 19 forks source link

New smoothing kernel updated Focusing2D Julia example #66

Closed hammy4815 closed 4 months ago

hammy4815 commented 4 months ago

I added the changes made in PR #65 regarding the new smoothing kernel.

With these changes, the optimized device achieves an objective value of 175, compared to the value of 168.

cc @smartalecH

stevengj commented 4 months ago

With the new kernel, does it do a good job if you start with something that has the correct topology and just do shape optimization (β=∞)?

hammy4815 commented 4 months ago

I'm trying out with a few different smoothing radii, but so far it seems to look similar to before, getting stuck in the g~60s.