[
  {
    "id": "ARR-2026-4GJP1XS78A95XS6H-v1",
    "type": "article",
    "title": "Exact Stabilization by Additive GKLS Control Is Impossible at a Leaky Boundary",
    "author": [
      {
        "literal": "Lluis Eriksson"
      }
    ],
    "issued": {
      "date-parts": [
        [
          2026,
          7,
          12
        ]
      ]
    },
    "publisher": "AIRR.SCIENCE",
    "archive": "AIRR.SCIENCE",
    "archive_location": "ARR-2026-4GJP1XS78A95XS6H",
    "version": "v1",
    "genre": "Archived manuscript",
    "URL": "https://airr.science/papers/ARR-2026-4GJP1XS78A95XS6H/versions/v1/",
    "note": "Archived manuscript; ARR-2026-4GJP1XS78A95XS6H v1. Not peer reviewed by AIRR. Historical import; not assessed by AIRR.",
    "abstract": "The tangent cone of quantum state space at a rank-deficient density matrix is known to have a positive exterior block, and recent work identifies all of its directions with Lindbladian velocities. We derive a quantitative stabilization consequence of this geometry. Let rho be a finite-dimensional target with support projector P, set Q = 1 - P, and let an uncontrolled GKLS generator L have outward support-leakage rate a = Tr[Q L(rho)] > 0. Every additive GKLS controller, including a bounded time-dependent controller, has its own nonnegative outward rate at rho and therefore cannot cancel a. Consequently, no finite-rate additive Markovian controller can keep the system exactly at the target. The conclusion persists for arbitrary finite-dimensional autonomous ancillas, provided the adverse system generator remains additive and local to the system.Under induced trace-norm bounds ||L|| <= M and ||K_t|| <= Gamma, we prove the dynamic corridor inequality limsup ||rho_t - rho||_1 >= 2a/(M + Gamma), without assuming convergence to a stationary state. An autonomous Poisson-reset generator supplies a matching inverse-rate upper bound, establishing the order-optimal minimax law Theta(Gamma^{-1}).For a soft-filter dual-rail SSH family, the microscopic leakage coefficient scales as exp[-(2m + q)ell + o(ell)], whereas the ideal logical disturbance scales as exp[-4m ell + O(1)]. We derive the exact controller-growth threshold g_min = max{0, 2m - q}. Thus, below the filter threshold q = 2m, retaining ideal logical accuracy requires an exponentially growing autonomous correction intensity."
  }
]