Cyclicity of Fredholm determinants of row-compact operators #
Let R be a complete ultrametric normed commutative ring with ‖1‖ = 1 and Tate scalars, and
I, J discrete index types. For an endomorphism w of the model space C₀(I, R), the
Fredholm determinant det (1 - T w) is the principal-minor series
w.matrixCoeff.charPowerSeries of its matrix.
This file proves that for continuous linear maps u : C₀(J, R) →L[R] C₀(I, R) and
v : C₀(I, R) →L[R] C₀(J, R), one of which is row compact, the determinants of the two
composites agree:
det (1 - T (u ∘ v)) = det (1 - T (v ∘ u))
(ContinuousLinearMap.IsRowCompact.charPowerSeries_matrixCoeff_comp_comm).
Both composites are endomorphisms, of C₀(I, R) and of C₀(J, R) respectively, and both are
row compact by the ideal property, so both series are entire
(Matrix.isRestricted_charPowerSeries). Consequently the Fredholm
determinant of a row-compact endomorphism is invariant under conjugation by any continuous linear
equivalence of model spaces, possibly between different index types
(ContinuousLinearMap.IsRowCompact.charPowerSeries_matrixCoeff_conj); no isometry is assumed.
The proof first treats an output truncation π_S ∘ u to a finite set S of coordinates. The
determinant of π_S ∘ u ∘ v is the polynomial det (1 - X (u ∘ v)_S) of the finite block on
S, while v ∘ π_S ∘ u factors through the finite free module R ^ S; its finite compressions
are products of rectangular blocks, to which the finite identity det (1 - X A B) = det (1 - X B A)
(Matrix.charpolyRev_mul_comm) applies. The general case follows by approximating u by its
output truncations, since each Fredholm coefficient is continuous in operator norm on row-compact
operators (ContinuousLinearMap.tendsto_charCoeff_matrixCoeff_of_tendsto).
Main results #
ContinuousLinearMap.tendsto_charCoeff_matrixCoeff_of_tendsto: the coefficients of the Fredholm determinant are continuous in operator norm on row-compact endomorphisms.ContinuousLinearMap.IsRowCompact.charPowerSeries_matrixCoeff_comp_comm: cyclicitydet (1 - T (u ∘ v)) = det (1 - T (v ∘ u))whenuis row compact.ContinuousLinearMap.IsRowCompact.charPowerSeries_matrixCoeff_conj: invariance under conjugation by a continuous linear equivalence.
References #
- J.-P. Serre, Endomorphismes complètement continus des espaces de Banach p-adiques, Publ. Math. IHÉS 12 (1962), 69–85, §§4–5.
- R. Coleman, p-adic Banach spaces and families of modular forms, Invent. Math. 127 (1997), 417–479, Appendix A.
The principal minors of each size of a row-compact endomorphism of a model space form a summable family, so the coefficients of its Fredholm determinant are genuine sums.
Continuity of Fredholm coefficients in operator norm. If row-compact endomorphisms
w a of a model space converge in operator norm to a row-compact endomorphism w₀, then each
coefficient of the Fredholm determinant of w a converges to the corresponding coefficient for
w₀.
Cyclicity of Fredholm determinants. If u : C₀(J, R) →L[R] C₀(I, R) is row compact and
v : C₀(I, R) →L[R] C₀(J, R) is any continuous linear map, the endomorphisms u ∘ v of
C₀(I, R) and v ∘ u of C₀(J, R) have the same Fredholm determinant:
det (1 - T (u ∘ v)) = det (1 - T (v ∘ u)). If instead v is row compact, apply this to v.
Cyclicity of Fredholm determinants for endomorphisms of one model space:
det (1 - T (u * v)) = det (1 - T (v * u)) when u is row compact.
Invariance under conjugation. The Fredholm determinant of a row-compact endomorphism u
of C₀(I, R) is unchanged by conjugation e ∘ u ∘ e⁻¹ by a continuous linear equivalence
e : C₀(I, R) ≃L[R] C₀(J, R), which need not be an isometry.