files / problems / hilbert16-limit-cycles / compute / q1 / README.md

1.2 KB · on GitHub

q1 — five lines backwards

Imagined end-states, then work backwards. No published H(n) moved.

Line Imagined claim Outcome
a-quadratic-five/ quadratic with 5 cycles dropped; fork: Shi origin is a weak focus of order 3, V3 = 35625/8
b-cubic-fourteen/ cubic with 14 cycles dropped; fork: van der Pol satisfies Liénard uniqueness (exactly 1 cycle)
c-chebyshev/ Chebyshev lift beats 2604.12883 Table 1 kept as replay; table matches; no missed factorization; §6 degree-11 field with 9 ovals
d-restricted-upper/ exact upper bound for a family kept: radial cubic has exactly 1 cycle; quadratic Hamiltonians have 0; Lotka–Volterra has 0 in the open first quadrant
e-bezout-bautin/ Bézout ceiling / Bautin L1 kept: one-step degree-m pullback ≤ m² sheets; L1 primitive polynomial. Full Bautin ideal dropped

Replay:

./run_all.sh

or from the problem compute folder, ./run_all.sh.

Certificates: a-quadratic-five/cert.json, b-cubic-fourteen/certificate.json, c-chebyshev/certs/, d-restricted-upper/certs/identities.json, e-bezout-bautin/*.json. Second languages: C (A), rustc (B–E), Lean 4.32.0 (e-bezout-bautin/AdjBezout.lean).