DiffusionControl / results /pendulum /constrained_smoke3.log
matCercola18's picture
Add files using upload-large-folder tool
929e74d verified
Raw History Blame Contribute Delete
1.93 kB
/home/cercola/matte_server/DiffusionControl/nonlinear/data_generation/pendulum_lmi.py:65: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
prob.solve(solver=cp.SCS, verbose=False)
/home/cercola/miniconda3/lib/python3.13/site-packages/numpy/_core/fromnumeric.py:3860: RuntimeWarning: Mean of empty slice.
return _methods._mean(a, axis=axis, dtype=dtype,
/home/cercola/miniconda3/lib/python3.13/site-packages/numpy/_core/_methods.py:145: RuntimeWarning: invalid value encountered in scalar divide
ret = ret.dtype.type(ret / rcount)
=== CONSTRAINED OC ===
Constrained optimal control on the pendulum (held-out N=10, budget E_max = 0.5× prior energy)
PRIOR (energy-unaware): median ∫u² = 18.557, median ISE = 0.0366 (uses 1.0× energy ⇒ violates the 0.5× budget)
STEERED toward min ISE s.t. ∫u² ≤ 0.5× prior (barrier λ=30.0):
satisfies budget: 0.00 median ∫u² = 17.723 median ISE = 0.0518 certified 1.00
ISE inflation vs prior (the cost of the budget): +42%
ISE REGRET vs per-plant constrained oracle: median +nan% (n=0)
ISE-vs-budget constrained-optimal frontier (steered):
ρ median E_max median ISE (sat) satisfied certified
0.35 6.495 nan 0.00 1.00
0.50 9.279 nan 0.00 1.00
0.70 12.990 nan 0.00 1.00
1.00 18.557 0.0365 1.00 1.00
INTERPRETATION: a non-quadratic, hard-CONSTRAINED optimal-control problem solved through
the differentiable non-linear pendulum — minimise tracking error subject to an energy
budget. The energy-unaware prior violates the budget; steering enforces it at the minimum
achievable tracking error (near the per-plant constrained optimum), staying certified;
tighter budgets cost more tracking error (the constrained-optimal frontier).