lq_regulator

LQ Regulator — optimal additive control (FRONTIERS V.2 · T1). Discounted linear-quadratic-Gaussian regulator on the recovered partial-adjustment plant: solves the discrete algebraic Riccati equation for the optimal additive-injection feedback, verifies certainty equivalence (the feedback is independent of the shock loading C — only the value constant moves), and reports the closed-loop spectral radius and the discounted-cost ranking optimal <= V.1-threshold < open. Reproduces FRONTIERS V.2 Theorem 1 on the stressed crisis calibration: the uncontrolled plant is non-stationary (rho 1.207, infinite discounted cost) and the optimal feedback stabilises it (closed-loop rho 0.424) at cost 540.2 < threshold 723.5 < infinity. The DARE is solved by a Riccati value-recursion plus a quadratically-convergent policy-iteration polish (DARE residual ~1e-15); it agrees with a Schur solver to ~4e-9, the DARE forward-error floor for this near-unstable plant. Bounded (n=26 fixture, sub-second). Gate: closed-loop rho / feedback / r(B) are STRONG.

method,theorem,control_target,control_target_gate,anchor,n,beta,q_ctrl,drift,rB_stress,rho_open,beta_rho2_open,open_loop_nonstationary,trace_P,F_norm2,rho_closed,closed_loop_stationary,certainty_equivalence_residual,d_C1,d_C2,J_optimal,J_threshold,J_open,rho_threshold,eta_safe,cost_ranking_holds,interpretation,gate,provenance

curl -s -X POST https://sys128-tower-1.tailac24de.ts.net/api/compute/run \
  -H "Content-Type: application/json" \
  -d '{"method":"lq_regulator","params":{}}'