def
definition
J_bit_normalized
show as:
view math explainer →
open explainer
Read the cached plain-language explainer.
open lean source
IndisputableMonolith.Constants.LambdaRecDerivation on GitHub at line 28.
browse module
All declarations in this module, on Recognition.
explainer page
depends on
used by
formal source
25/-! ## The Balance Condition -/
26
27/-- Bit cost of one recognition event (normalized). -/
28noncomputable def J_bit_normalized : ℝ := 1
29
30/-- Curvature cost of embedding one recognition token at scale λ.
31 Derived from |κ| = 4 curvature quanta on Q₃, Gauss-Bonnet
32 normalization on S² with χ = 2, and bounding area 4πλ².
33 J_curv = (|κ|/(2χ)) · (A/(2π)) = (4/4) · (4πλ²/(2π)) = 2λ². -/
34noncomputable def J_curv (lambda : ℝ) : ℝ := 2 * lambda ^ 2
35
36/-- Total cost functional. -/
37noncomputable def totalCost (lambda : ℝ) : ℝ :=
38 J_bit_normalized + J_curv lambda
39
40/-- Balance residual: vanishes at the optimal scale. -/
41noncomputable def balanceResidual (lambda : ℝ) : ℝ :=
42 J_curv lambda - J_bit_normalized
43
44/-- The balance point in RS-native dimensionless units. -/
45noncomputable def lambda_0 : ℝ := 1 / Real.sqrt 2
46
47lemma lambda_0_pos : 0 < lambda_0 := by
48 unfold lambda_0
49 apply div_pos one_pos
50 exact Real.sqrt_pos.mpr (by norm_num : (0 : ℝ) < 2)
51
52/-- lambda_0² = 1/2. -/
53lemma lambda_0_sq : lambda_0 ^ 2 = 1 / 2 := by
54 unfold lambda_0
55 rw [div_pow]
56 have h2 : (0 : ℝ) ≤ 2 := by norm_num
57 rw [Real.sq_sqrt h2]
58 norm_num