You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Proposal 4: structural genes, so the loop can change the structure #7
Closes: the "evolving its hardware" part of the vision (docs/VISION.md in #2).
The gap
The loop changes three thermal parameters: the thermal store, the liquid capture fraction and the wall U-value. The eleven certified structural margins (punching collar, storey drift, roof uplift and so on) are constants in it, because the aska-physical verifiers that compute them take no inputs. So today "hardware" means three numbers in a thermal model and no member of the structure.
The system
Each aska-physical verifier gets a parameter file with its design inputs and an --inputs option. The loop can then ask a real question, such as "what are the margins if the floor load is 5% higher", and get the answer from the upstream verifier.
First slice
funi-frame's verify.py reads its roof and wall inputs from a file.
The loop gains one gene, the rack floor load, and re-runs funi-frame and funi-floors for each candidate.
Done when a proposal that raises the floor load past the punching collar's capacity is refused, with the margin computed by the upstream verifier and not by a scaling rule. The collar is the tightest certified margin in the building at 80.1%, so this is the first structural limit the loop will meet.
Rests on
funi-frame, funi-floors.
Also fixes
Six values are copied by hand between funi-frame, funi-floors and venturi-collars today (optifab checks that they agree). With parameter files they are read, not copied.
Filed by an AI agent (Claude Code) working for the owner of aska-physical.
Activity