Finite parity pencils converge monotonically to the full generalized Rayleigh budget interval.
Theorem 1.1 (Finite Hermitian pencils bound and approximate the full budget interval).
∀ E v e n T es t ∈ T y p e , O dd T es t ∈ T y p e , e v e n D im ∈ Nat ( ) → Nat ( ) , o dd D im ∈ Nat ( ) → Nat ( ) , e v e n B a se ∈ DependentMap ( N : Nat ( ) , Matrix ( Fin ( apply ( e v e n D im , N ) ) , Fin ( apply ( e v e n D im , N ) ) , Complex ( ) ) ) , o dd B a se ∈ DependentMap ( N : Nat ( ) , Matrix ( Fin ( apply ( o dd D im , N ) ) , Fin ( apply ( o dd D im , N ) ) , Complex ( ) ) ) , e v e n B o u n d a ry ∈ DependentMap ( N : Nat ( ) , Fin ( apply ( e v e n D im , N ) ) → Complex ( ) ) , o dd B o u n d a ry ∈ DependentMap ( N : Nat ( ) , Fin ( apply ( o dd D im , N ) ) → Complex ( ) ) , f u llE v e n B a se ∈ E v e n T es t → Real ( ) , f u llE v e n B o u n d a ry ∈ E v e n T es t → Complex ( ) , f u llO dd B a se ∈ O dd T es t → Real ( ) , f u llO dd B o u n d a ry ∈ O dd T es t → Complex ( ) , re f ere n ce B u d g e t ∈ Real ( ) , let f ini t e E v e n Q u o t i e n t s = N : Nat ( ) ↦ { ∃ x ∈ Fin ( apply ( e v e n D im , N ) ) → Complex ( ) , dot ( star ( apply ( e v e n B o u n d a ry , N ) ) , x ) = 0 ∧ q = div ( neg ( Re ( dot ( star ( x ) , mulVec ( apply ( e v e n B a se , N ) , x ) ) ) ) , normSq ( dot ( star ( apply ( e v e n B o u n d a ry , N ) ) , x ) ) ) ∣ q ∈ Real ( ) } , let f ini t e O dd Q u o t i e n t s = N : Nat ( ) ↦ { ∃ x ∈ Fin ( apply ( o dd D im , N ) ) → Complex ( ) , dot ( star ( apply ( o dd B o u n d a ry , N ) ) , x ) = 0 ∧ q = div ( Re ( dot ( star ( x ) , mulVec ( apply ( o dd B a se , N ) , x ) ) ) , normSq ( dot ( star ( apply ( o dd B o u n d a ry , N ) ) , x ) ) ) ∣ q ∈ Real ( ) } , let f u llE v e n Q u o t i e n t s = { ∃ x ∈ E v e n T es t , apply ( f u llE v e n B o u n d a ry , x ) = 0 ∧ q = div ( neg ( apply ( f u llE v e n B a se , x ) ) , normSq ( apply ( f u llE v e n B o u n d a ry , x ) ) ) ∣ q ∈ Real ( ) } , let f u llO dd Q u o t i e n t s = { ∃ x ∈ O dd T es t , apply ( f u llO dd B o u n d a ry , x ) = 0 ∧ q = div ( apply ( f u llO dd B a se , x ) , normSq ( apply ( f u llO dd B o u n d a ry , x ) ) ) ∣ q ∈ Real ( ) } , let f ini t e L o w er = N : Nat ( ) ↦ add ( re f ere n ce B u d g e t , sSup ( apply ( f ini t e E v e n Q u o t i e n t s , N ) ) ) , let f ini t e U pp er = N : Nat ( ) ↦ add ( re f ere n ce B u d g e t , sInf ( apply ( f ini t e O dd Q u o t i e n t s , N ) ) ) , let f u ll L o w er = add ( re f ere n ce B u d g e t , sSup ( f u llE v e n Q u o t i e n t s ) ) , let f u ll U pp er = add ( re f ere n ce B u d g e t , sInf ( f u llO dd Q u o t i e n t s ) ) , let e v e n P e n c i l = N : Nat ( ) , R : Real ( ) ↦ add ( apply ( e v e n B a se , N ) , smul ( ofReal ( sub ( R , re f ere n ce B u d g e t ) ) , vecMulVec ( apply ( e v e n B o u n d a ry , N ) , star ( apply ( e v e n B o u n d a ry , N ) ) ) ) ) , let o dd P e n c i l = N : Nat ( ) , R : Real ( ) ↦ sub ( apply ( o dd B a se , N ) , smul ( ofReal ( sub ( R , re f ere n ce B u d g e t ) ) , vecMulVec ( apply ( o dd B o u n d a ry , N ) , star ( apply ( o dd B o u n d a ry , N ) ) ) ) ) , let f e a s ib l e = N : Nat ( ) , R : Real ( ) ↦ PosSemidef ( applyTwo ( e v e n P e n c i l , N , R ) ) ∧ PosSemidef ( applyTwo ( o dd P e n c i l , N , R ) ) , ( ( ∀ N ∈ Nat ( ) , IsHermitian ( apply ( e v e n B a se , N ) ) ) ∧ ( ( ∀ N ∈ Nat ( ) , IsHermitian ( apply ( o dd B a se , N ) ) ) ∧ ( ( ∀ N ∈ Nat ( ) , x ∈ Fin ( apply ( e v e n D im , N ) ) → Complex ( ) , dot ( star ( apply ( e v e n B o u n d a ry , N ) ) , x ) = 0 ⇒ 0 ≤ Re ( dot ( star ( x ) , mulVec ( apply ( e v e n B a se , N ) , x ) ) ) ) ∧ ( ( ∀ N ∈ Nat ( ) , x ∈ Fin ( apply ( o dd D im , N ) ) → Complex ( ) , dot ( star ( apply ( o dd B o u n d a ry , N ) ) , x ) = 0 ⇒ 0 ≤ Re ( dot ( star ( x ) , mulVec ( apply ( o dd B a se , N ) , x ) ) ) ) ∧ ( ( ∀ N ∈ Nat ( ) , ∃ x ∈ Fin ( apply ( e v e n D im , N ) ) → Complex ( ) , dot ( star ( apply ( e v e n B o u n d a ry , N ) ) , x ) = 0 ) ∧ ( ( ∀ N ∈ Nat ( ) , ∃ x ∈ Fin ( apply ( o dd D im , N ) ) → Complex ( ) , dot ( star ( apply ( o dd B o u n d a ry , N ) ) , x ) = 0 ) ∧ ( ( ∀ N ∈ Nat ( ) , BddAbove ( apply ( f ini t e E v e n Q u o t i e n t s , N ) ) ) ∧ ( ( ∀ N ∈ Nat ( ) , BddBelow ( apply ( f ini t e O dd Q u o t i e n t s , N ) ) ) ∧ ( ( ∃ x ∈ E v e n T es t , apply ( f u llE v e n B o u n d a ry , x ) = 0 ) ∧ ( ( ∃ x ∈ O dd T es t , apply ( f u llO dd B o u n d a ry , x ) = 0 ) ∧ ( BddAbove ( f u llE v e n Q u o t i e n t s ) ∧ ( BddBelow ( f u llO dd Q u o t i e n t s ) ∧ ( ( ∀ N ∈ Nat ( ) , Subset ( apply ( f ini t e E v e n Q u o t i e n t s , N ) , apply ( f ini t e E v e n Q u o t i e n t s , add ( N , 1 ) ) ) ) ∧ ( ( ∀ N ∈ Nat ( ) , Subset ( apply ( f ini t e O dd Q u o t i e n t s , N ) , apply ( f ini t e O dd Q u o t i e n t s , add ( N , 1 ) ) ) ) ∧ ( ( ∀ N ∈ Nat ( ) , Subset ( apply ( f ini t e E v e n Q u o t i e n t s , N ) , f u llE v e n Q u o t i e n t s ) ) ∧ ( ( ∀ N ∈ Nat ( ) , Subset ( apply ( f ini t e O dd Q u o t i e n t s , N ) , f u llO dd Q u o t i e n t s ) ) ∧ ( ( ∀ q ∈ Real ( ) , e p s i l o n ∈ Real ( ) , ( Mem ( q , f u llE v e n Q u o t i e n t s ) ∧ 0 < e p s i l o n ) ⇒ ( ∃ N ∈ Nat ( ) , qN ∈ Real ( ) , Mem ( qN , apply ( f ini t e E v e n Q u o t i e n t s , N ) ) ∧ sub ( q , e p s i l o n ) < qN ) ) ∧ ( ∀ q ∈ Real ( ) , e p s i l o n ∈ Real ( ) , ( Mem ( q , f u llO dd Q u o t i e n t s ) ∧ 0 < e p s i l o n ) ⇒ ( ∃ N ∈ Nat ( ) , qN ∈ Real ( ) , Mem ( qN , apply ( f ini t e O dd Q u o t i e n t s , N ) ) ∧ qN < add ( q , e p s i l o n ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ⇒ ( ( ∀ N ∈ Nat ( ) , R ∈ Real ( ) , IsHermitian ( applyTwo ( e v e n P e n c i l , N , R ) ) ∧ IsHermitian ( applyTwo ( o dd P e n c i l , N , R ) ) ) ∧ ( ( ∀ N ∈ Nat ( ) , R ∈ Real ( ) , Iff ( PosSemidef ( applyTwo ( e v e n P e n c i l , N , R ) ) , apply ( f ini t e L o w er , N ) ≤ R ) ) ∧ ( ( ∀ N ∈ Nat ( ) , R ∈ Real ( ) , Iff ( PosSemidef ( applyTwo ( o dd P e n c i l , N , R ) ) , R ≤ apply ( f ini t e U pp er , N ) ) ) ∧ ( Monotone ( f ini t e L o w er ) ∧ ( Antitone ( f ini t e U pp er ) ∧ ( Tendsto ( f ini t e L o w er , a tT o p , nhds ( f u ll L o w er ) ) ∧ ( Tendsto ( f ini t e U pp er , a tT o p , nhds ( f u ll U pp er ) ) ∧ ( ( ∀ N ∈ Nat ( ) , R ∈ Real ( ) , Iff ( applyTwo ( f e a s ib l e , N , R ) , MemIcc ( R , apply ( f ini t e L o w er , N ) , apply ( f ini t e U pp er , N ) ) ) ) ∧ ( ∀ N ∈ Nat ( ) , apply ( f ini t e U pp er , N ) < apply ( f ini t e L o w er , N ) ⇒ ( ( ∀ R ∈ Real ( ) , PosSemidef ( applyTwo ( e v e n P e n c i l , N , R ) ) ⇒ apply ( f ini t e L o w er , N ) ≤ R ) ∧ ( ( ∀ R ∈ Real ( ) , PosSemidef ( applyTwo ( o dd P e n c i l , N , R ) ) ⇒ R ≤ apply ( f ini t e U pp er , N ) ) ∧ Not ( ∃ R ∈ Real ( ) , applyTwo ( f e a s ib l e , N , R ) ) ) ) ) ) ) ) ) ) ) ) )
Proof. Machine-checked in Lean as D5/S3/Weil/Budget/FiniteWeylBudgetBounds.finite_weyl_budget_bounds (✓ std3). ∎
Source. Repository-derived.
Commentary.
The displayed statement constructs both finite quotient families, their endpoints, the two rank-one pencils, and finite feasibility directly from the matrix and boundary data.
Kernel positivity handles zero boundary pairings. Nested finite quotient sets and one-sided approximation of every full quotient express the Galerkin density hypothesis without assuming endpoint convergence.
Each finite pencil is positive semidefinite exactly on its corresponding budget ray. The lower endpoints increase, the upper endpoints decrease, both converge to the full interval, and a crossed finite interval supplies incompatible lower and upper requirements.