AFTD Auto-Formalizing Theoretical Domains

parallel_mw_balancing_strict_core

theoremverified

The stated transfer and numerator bounds imply the strict cross-multiplied balancing inequality when d and N are positive.

Statement

theorem parallel_mw_balancing_strict_core (s t A B d N : ℝ)
    (hs : 0 < s) (ht : 2*s ≤ t) (hA : 0 ≤ A) (hB : A ≤ B)
    (hd : 0 < d) (hN : 0 < N)
    (hcases : (d = 1 ∧ B = 2*A) ∨ (d = 2 ∧ B = 4*A) ∨
      (d = 3 ∧ B = 8*A) ∨ 4 ≤ d)
    (hbound : B*(s+t) ≤ 2*N) :
    2*(B-A)*(s*d+t)^2 < N*(2*(s*d+t)*d+d^2)
Source
Algebraic reduction for parallel_mw_ratio_decreasing_triple and OP-670.
Verified
02 Nov 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 1
Used by 1

The proof, explained

The proof bounds the left-hand side 2(B - A)(sd + t)^2 using the transfer bound and the budget condition B(s + t) ≤ 2N. First, the transfer bound gives 2(B - A)(sd + t) ≤ Bd(s + t). Multiplying this by the positive factor sd + t and chaining with the scaled budget bound Bd(sd + t)(s + t) ≤ 2Nd(sd + t) bounds the left-hand side by 2Nd(sd + t). Finally, adding the strictly positive remainder term Nd^2 strictly increases the bound to N(2(sd + t)d + d^2), which nonlinear arithmetic verifies.

  1. 2 * (B - A) * (s * d + t) ≤ B * d * (s + t)

    Apply the transfer lemma `parallel_mw_balancing_transfer_bound` using the non-negativity of s (from s > 0), the ratio bound 2s ≤ t, the order 0 ≤ A ≤ B, and the case condition on d and B.

    line 8

    uses parallel_mw_balancing_transfer_bound

      have htr := parallel_mw_balancing_transfer_bound s t A B d hs.le ht hA hB hcases
  2. 0 < t and 0 < s * d + t

    Because 0 < s and 2s ≤ t, linear arithmetic gives 0 < t. Then, since s, d, and t are all strictly positive, their combination s * d + t is strictly positive by positivity reasoning.

    lines 9–10

      have htpos : 0 < t := by linarith
      have hu : 0 < s*d+t := by positivity
  3. The products (s * d + t) * (B * d * (s + t) - 2 * (B - A) * (s * d + t)) and d * (s * d + t) * (2 * N - B * (s + t)) are non-negative, and N * d^2 is strictly positive.

    Rewriting the transfer bound and the budget bound as non-negative differences via `sub_nonneg.mpr` and multiplying by the non-negative quantities s * d + t and d * (s * d + t) yields two non-negative products via `mul_nonneg`. Furthermore, since d > 0, d^2 > 0 by `sq_pos_of_pos`, and since N > 0, the product N * d^2 is strictly positive by `mul_pos`.

    lines 11–13

    uses mul_nonneg, sub_nonneg.mpr, mul_pos, sq_pos_of_pos

      have h1 := mul_nonneg hu.le (sub_nonneg.mpr htr)
      have h2 := mul_nonneg (mul_nonneg hd.le hu.le) (sub_nonneg.mpr hbound)
      have h3 := mul_pos hN (sq_pos_of_pos hd)
  4. 2 * (B - A) * (s * d + t)^2 < N * (2 * (s * d + t) * d + d^2) automation

    The target strict inequality follows because the right-hand side minus the left-hand side equals the sum of the two non-negative products and the strictly positive term N * d^2; nonlinear arithmetic (`nlinarith`) closes the goal.

    line 14

      nlinarith
parallel_mw_balancing_transfer_bound
If 0 ≤ s, 2s ≤ t, 0 ≤ A ≤ B, and (d, B) satisfy the case condition, then 2(B - A)(sd + t) ≤ Bd(s + t).
mul_nonneg
The product of two non-negative numbers is non-negative.
sub_nonneg.mpr
If b ≤ a, then 0 ≤ a - b.
mul_pos
The product of two strictly positive numbers is strictly positive.
sq_pos_of_pos
The square of a strictly positive number is strictly positive.

Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

theorem parallel_mw_balancing_strict_core (s t A B d N : ℝ)    (hs : 0 < s) (ht : 2*s ≤ t) (hA : 0 ≤ A) (hB : A ≤ B)    (hd : 0 < d) (hN : 0 < N)    (hcases : (d = 1 ∧ B = 2*A) ∨ (d = 2 ∧ B = 4*A) ∨      (d = 3 ∧ B = 8*A) ∨ 4 ≤ d)    (hbound : B*(s+t) ≤ 2*N) :    2*(B-A)*(s*d+t)^2 < N*(2*(s*d+t)*d+d^2) := by  have htr := parallel_mw_balancing_transfer_bound s t A B d hs.le ht hA hB hcases  have htpos : 0 < t := by linarith  have hu : 0 < s*d+t := by positivity  have h1 := mul_nonneg hu.le (sub_nonneg.mpr htr)  have h2 := mul_nonneg (mul_nonneg hd.le hu.le) (sub_nonneg.mpr hbound)  have h3 := mul_pos hN (sq_pos_of_pos hd)  nlinarith
See it in the graphDiscuss this resultChallenge it