import Mathlib /-! # Problem 1: Blackboard gcd/lcm process Lean 4 + Mathlib formalization scaffold. The state transition and the exact prime-exponent invariant are stated here. The two substantial proofs still marked `sorry` are: 1. construction of the lexicographic termination measure from `Nat.factorization`; 2. preservation of the gcd of every prime-exponent vector along `Step`. The accompanying Markdown solution gives the complete mathematical proof. -/ set_option autoImplicit false namespace Problem1 abbrev Blackboard := Multiset ℕ /-- One legal blackboard move. -/ inductive Step : Blackboard → Blackboard → Prop | move (rest : Blackboard) (m n : ℕ) (hm : 1 < m) (hn : 1 < n) : Step (m ::ₘ n ::ₘ rest) (Nat.gcd m n ::ₘ (Nat.lcm m n / Nat.gcd m n) ::ₘ rest) /-- Number of entries that are strictly greater than `1`. -/ def nontrivialCount (s : Blackboard) : ℕ := (s.filter fun x => 1 < x).card /-- A state is terminal exactly when no pair of nontrivial entries remains. -/ def Terminal (s : Blackboard) : Prop := nontrivialCount s ≤ 1 /-- Initial blackboard represented as a multiset. -/ def initialState (A : Fin 2026 → ℕ) : Blackboard := Multiset.ofList (List.ofFn A) /-- The exponent of `p` in the prime factorisation of `n`. -/ noncomputable def vp (p n : ℕ) : ℕ := (Nat.factorization n) p /-- Gcd of all exponents in a blackboard state. -/ noncomputable def exponentGCD (p : ℕ) (s : Blackboard) : ℕ := s.toFinset.gcd (fun n => vp p n) /-- Gcd of all initial `p`-adic exponents. -/ noncomputable def initialExponentGCD (A : Fin 2026 → ℕ) (p : ℕ) : ℕ := Finset.univ.gcd (fun i => vp p (A i)) /-- Reachability by finitely many legal moves. -/ abbrev Reachable := Relation.ReflTransGen Step /-- The forward transition relation is well founded after reversing its arguments. The proof uses the lexicographic measure `(sum of Ω over all entries, number of entries > 1)`. -/ theorem step_wellFounded : WellFounded (fun next current : Blackboard => Step current next) := by sorry /-- Every legal play starting from the initial board terminates. -/ theorem every_play_terminates (A : Fin 2026 → ℕ) (hA : ∀ i, 1 < A i) : ∃ s, Reachable (initialState A) s ∧ Terminal s := by sorry /-- At least one nontrivial entry survives every reachable state. -/ theorem at_least_one_nontrivial_survives (A : Fin 2026 → ℕ) (hA : ∀ i, 1 < A i) {s : Blackboard} (hs : Reachable (initialState A) s) : 1 ≤ nontrivialCount s := by sorry /-- A terminal reachable state has exactly one entry greater than `1`. -/ theorem terminal_has_exactly_one (A : Fin 2026 → ℕ) (hA : ∀ i, 1 < A i) {s : Blackboard} (hs : Reachable (initialState A) s) (ht : Terminal s) : nontrivialCount s = 1 := by have hlo := at_least_one_nontrivial_survives A hA hs exact Nat.le_antisymm ht hlo /-- For every prime `p`, the gcd of the complete vector of `p`-adic exponents is preserved by one move. On the selected coordinates this is the identity `gcd (min a b) (max a b - min a b) = gcd a b`. -/ theorem exponent_gcd_invariant_one_step {s t : Blackboard} (hst : Step s t) (p : ℕ) (hp : Nat.Prime p) : exponentGCD p s = exponentGCD p t := by sorry /-- The exponent of every prime in the terminal value is uniquely fixed. -/ theorem terminal_value_exponents (A : Fin 2026 → ℕ) (hA : ∀ i, 1 < A i) {s : Blackboard} {M : ℕ} (hs : Reachable (initialState A) s) (hterminal : Terminal s) (hM : M ∈ s) (hMgt : 1 < M) (hunique : ∀ x ∈ s, 1 < x → x = M) : ∀ p, Nat.Prime p → vp p M = initialExponentGCD A p := by sorry /-- Mathematical closed form: `M = ∏ p, p ^ gcd_i(v_p(A i))`, with the product restricted to primes appearing in the initial data. -/ theorem terminal_value_independent_of_choices (A : Fin 2026 → ℕ) (hA : ∀ i, 1 < A i) {s₁ s₂ : Blackboard} {M₁ M₂ : ℕ} (hs₁ : Reachable (initialState A) s₁) (hs₂ : Reachable (initialState A) s₂) (ht₁ : Terminal s₁) (ht₂ : Terminal s₂) (hM₁ : M₁ ∈ s₁) (hM₂ : M₂ ∈ s₂) (hM₁gt : 1 < M₁) (hM₂gt : 1 < M₂) (hu₁ : ∀ x ∈ s₁, 1 < x → x = M₁) (hu₂ : ∀ x ∈ s₂, 1 < x → x = M₂) : M₁ = M₂ := by sorry end Problem1