Jump to content

Nuclear operator

From Wikipedia, the free encyclopedia

In mathematics, nuclear operators are an important class of linear operators introduced by Alexander Grothendieck in his doctoral dissertation[1]. Nuclear operators are intimately tied to the projective tensor product of two topological vector spaces (TVSs).[2]

Preliminaries and notation

[edit]

Throughout let X,Y, and Z be topological vector spaces (TVSs) and L : X → Y be a linear operator (no assumption of continuity is made unless otherwise stated).

  • The projective tensor product of two locally convex TVSs X and Y is denoted by X⊗πY and the completion of this space will be denoted by X⊗^πY.
  • L : X → Y is a topological homomorphism or homomorphism, if it is linear, continuous, and L:X→Im⁡L is an open map, where Im⁡L, the image of L, has the subspace topology induced by Y.
    • If S is a subspace of X then both the quotient map X → X/S and the canonical injection S → X are homomorphisms.
  • The set of continuous linear maps X → Z (resp. continuous bilinear maps X×Y→Z) will be denoted by L(X, Z) (resp. B(X, Y; Z)) where if Z is the underlying scalar field then we may instead write L(X) (resp. B(X, Y)).
  • Any linear map L:X→Y can be canonically decomposed as follows: X→X/ker⁡L→L0Im⁡L→Y where L0(x+ker⁡L):=L(x) defines a bijection called the canonical bijection associated with L.
  • X* or X′ will denote the continuous dual space of X.
    • To increase the clarity of the exposition, we use the common convention of writing elements of X′ with a prime following the symbol (e.g. x′ denotes an element of X′ and not, say, a derivative and the variables x and x′ need not be related in any way).
  • X# will denote the algebraic dual space of X (which is the vector space of all linear functionals on X, whether continuous or not).
  • A linear map L : H → H from a Hilbert space into itself is called positive if ⟨L(x),x⟩≥0 for every x∈H. In this case, there is a unique positive map r : H → H, called the square-root of L, such that L=r∘r.[3]
    • If L:H1→H2 is any continuous linear map between Hilbert spaces, then L∗∘L is always positive. Now let R : H → H denote its positive square-root, which is called the absolute value of L. Define U:H1→H2 first on Im⁡R by setting U(x)=L(x) for x=R(x1)∈Im⁡R and extending U continuously to Im⁡R‾, and then define U on ker⁡R by setting U(x)=0 for x∈ker⁡R and extend this map linearly to all of H1. The map U|Im⁡R:Im⁡R→Im⁡L is a surjective isometry and L=U∘R.
  • A linear map Λ:X→Y is called compact or completely continuous if there is a neighborhood U of the origin in X such that Λ(U) is precompact in Y.[4]

In a Hilbert space, positive compact linear operators, say L : H → H have a simple spectral decomposition discovered at the beginning of the 20th century by Fredholm and F. Riesz:[5]

There is a sequence of positive numbers, decreasing and either finite or else converging to 0, r1>r2>⋯>rk>⋯ and a sequence of nonzero finite dimensional subspaces Vi of H (i = 1, 2, …) with the following properties: (1) the subspaces Vi are pairwise orthogonal; (2) for every i and every x∈Vi, L(x)=rix; and (3) the orthogonal of the subspace spanned by ⋃iVi is equal to the kernel of L.[5]

Notation for topologies

[edit]

A canonical tensor product as a subspace of the dual of Bi(X, Y)

[edit]

Let X and Y be vector spaces (no topology is needed yet) and let Bi(X, Y) be the space of all bilinear maps defined on X×Y and going into the underlying scalar field.

For every (x,y)∈X×Y, let χ(x,y) be the canonical linear form on Bi(X, Y) defined by χ(x,y)(u):=u(x,y) for every u ∈ Bi(X, Y). This induces a canonical map χ:X×Y→Bi(X,Y)# defined by χ(x,y):=χ(x,y), where Bi(X,Y)# denotes the algebraic dual of Bi(X, Y). If we denote the span of the range of 𝜒 by X ⊗ Y then it can be shown that X ⊗ Y together with 𝜒 forms a tensor product of X and Y (where x ⊗ y := 𝜒(x, y)). This gives us a canonical tensor product of X and Y.

If Z is any other vector space then the mapping Li(X ⊗ Y; Z) → Bi(X, Y; Z) given by u ↦ u ∘ 𝜒 is an isomorphism of vector spaces. In particular, this allows us to identify the algebraic dual of X ⊗ Y with the space of bilinear forms on X × Y.[6] Moreover, if X and Y are locally convex topological vector spaces (TVSs) and if X ⊗ Y is given the π-topology then for every locally convex TVS Z, this map restricts to a vector space isomorphism L(X⊗πY;Z)→B(X,Y;Z) from the space of continuous linear mappings onto the space of continuous bilinear mappings.[7] In particular, the continuous dual of X ⊗ Y can be canonically identified with the space B(X, Y) of continuous bilinear forms on X × Y; furthermore, under this identification the equicontinuous subsets of B(X, Y) are the same as the equicontinuous subsets of (X⊗πY)′.[7]

Nuclear operators between Banach spaces

[edit]

There is a canonical vector space embedding I:X′⊗Y→L(X;Y) defined by sending z:=∑inxi′⊗yi to the map x↦∑inxi′(x)yi.

Assuming that X and Y are Banach spaces, then the map I:X'b⊗πY→Lb(X;Y) has norm 1 (to see that the norm is ≤1, note that

‖I(z)‖=sup‖x‖≤1‖I(z)(x)‖=sup‖x‖≤1‖∑i=1nxi′(x)yi‖≤sup‖x‖≤1∑i=1n‖xi′‖‖x‖‖yi‖≤∑i=1n‖xi′‖‖yi‖, so that ‖I(z)‖≤‖z‖π).

Thus I has a continuous extension to a map Î:X'b⊗^πY→Lb(X;Y). It is known that this map is not necessarily injective.[8] The range of this map is denoted by L1(X;Y) and its elements are called nuclear operators.[9]

L1(X;Y) is TVS-isomorphic to (X'b⊗^πY)/ker⁡Î and the norm on this quotient space, when transferred to elements of L1(X;Y) via the induced map Î:(X'b⊗^πY)/ker⁡Î→L1(X;Y), is called the trace-norm and is denoted by ‖⋅‖Tr. Explicitly,[clarification needed explicitly or especially?] if T:X→Y is a nuclear operator then ‖T‖Tr:=infz∈Î−1(T)‖z‖π.

Characterization

[edit]

Suppose that X and Y are Banach spaces and that N:X→Y is a continuous linear operator.

  • The following are equivalent:
    1. N:X→Y is nuclear.
    2. There exists a sequence (xi′)i=1∞ in the closed unit ball of X′, a sequence (yi)i=1∞ in the closed unit ball of Y, and a complex sequence (ci)i=1∞ such that ∑i=1∞|ci|<∞ and N is equal to the mapping:[10] N(x)=∑i=1∞cix'i(x)yi for all x∈X. Furthermore, the trace-norm ‖N‖Tr is equal to the infimum of the numbers ∑i=1∞|ci| over the set of all representations of N as such a series.[10]
  • If Y is reflexive then N:X→Y is a nuclear if and only if tN:Y'b→X'b is nuclear, in which case ‖tN‖Tr=‖N‖Tr. [11]

Properties

[edit]

Let X and Y be Banach spaces and let N:X→Y be a continuous linear operator.

  • If N:X→Y is a nuclear map then its transpose tN:Y'b→X'b is a continuous nuclear map (when the dual spaces carry their strong dual topologies) and ‖tN‖Tr≤‖N‖Tr.[12]

Nuclear operators between Hilbert spaces

[edit]

Nuclear automorphisms of a Hilbert space are called trace class operators.

Let X and Y be Hilbert spaces and let N : X → Y be a continuous linear map. Suppose that N=UR where R : X → X is the square-root of N∗N and U : X → Y is such that U|Im⁡R:Im⁡R→Im⁡N is a surjective isometry. Then N is a nuclear map if and only if R is a nuclear map; hence, to study nuclear maps between Hilbert spaces it suffices to restrict one's attention to positive self-adjoint operators R.[13]

Characterizations

[edit]

Let X and Y be Hilbert spaces and let N : X → Y be a continuous linear map whose absolute value is R : X → X. The following are equivalent:

  1. N : X → Y is nuclear.
  2. R : X → X is nuclear.[14]
  3. R : X → X is compact and Tr⁡R is finite, in which case Tr⁡R=‖N‖Tr.[14]
    • Here, Tr⁡R is the trace of R and it is defined as follows: Since R is a continuous compact positive operator, there exists a (possibly finite) sequence λ1>λ2>⋯ of positive numbers with corresponding non-trivial finite-dimensional and mutually orthogonal vector spaces V1,V2,… such that the orthogonal (in H) of span⁡(V1∪V2∪⋯) is equal to ker⁡R (and hence also to ker⁡N) and for all k, R(x)=λkx for all x∈Vk; the trace is defined as Tr⁡R:=∑kλkdim⁡Vk.
  4. tN:Y'b→X'b is nuclear, in which case ‖tN‖Tr=‖N‖Tr. [11]
  5. There are two orthogonal sequences (xi)i=1∞ in X and (yi)i=1∞ in Y, and a sequence (λi)i=1∞ in ℓ1 such that for all x∈X, N(x)=∑iλi⟨x,xi⟩yi.[14]
  6. N : X → Y is an integral map.[15]

Nuclear operators between locally convex spaces

[edit]

Suppose that U is a convex balanced closed neighborhood of the origin in X and B is a convex balanced bounded Banach disk in Y with both X and Y locally convex spaces. Let pU(x)=infr>0,x∈rUr and let π:X→X/pU−1(0) be the canonical projection. One can define the auxiliary Banach space X̂U with the canonical map π̂U:X→X̂U whose image, X/pU−1(0), is dense in X̂U as well as the auxiliary space FB=span⁡B normed by pB(y)=infr>0,y∈rBr and with a canonical map ι:FB→F being the (continuous) canonical injection. Given any continuous linear map T:X̂U→YB one obtains through composition the continuous linear map π̂U∘T∘ι:X→Y; thus we have an injection L(X̂U;YB)→L(X;Y) and we henceforth use this map to identify L(X̂U;YB) as a subspace of L(X;Y).[9]

Definition: Let X and Y be Hausdorff locally convex spaces. The union of all L1(X̂U;YB) as U ranges over all closed convex balanced neighborhoods of the origin in X and B ranges over all bounded Banach disks in Y, is denoted by L1(X;Y) and its elements are call nuclear mappings of X into Y.[9]

When X and Y are Banach spaces, then this new definition of nuclear mapping is consistent with the original one given for the special case where X and Y are Banach spaces.

Sufficient conditions for nuclearity

[edit]
  • Let W, X, Y, and Z be Hausdorff locally convex spaces, N:X→Y a nuclear map, and M:W→X and P:Y→Z be continuous linear maps. Then N∘M:W→Y, P∘N:X→Z, and P∘N∘M:W→Z are nuclear and if in addition W, X, Y, and Z are all Banach spaces then ‖P∘N∘M‖Tr≤‖P‖‖N‖Tr‖‖M‖.[16][17]
  • If N:X→Y is a nuclear map between two Hausdorff locally convex spaces, then its transpose tN:Y'b→X'b is a continuous nuclear map (when the dual spaces carry their strong dual topologies).[4]
    • If in addition X and Y are Banach spaces, then ‖tN‖Tr≤‖N‖Tr.[11]
  • If N:X→Y is a nuclear map between two Hausdorff locally convex spaces and if X̂ is a completion of X, then the unique continuous extension N̂:X̂→Y of N is nuclear.[17]

Characterizations

[edit]

Let X and Y be Hausdorff locally convex spaces and let N:X→Y be a continuous linear operator.

  • The following are equivalent:
    1. N:X→Y is nuclear.
    2. (Definition) There exists a convex balanced neighborhood U of the origin in X and a bounded Banach disk B in Y such that N(U)⊆B and the induced map N‾0:X̂U→YB is nuclear, where N‾0 is the unique continuous extension of N0:XU→YB, which is the unique map satisfying N=InB∘N0∘πU where InB:YB→Y is the natural inclusion and πU:X→X/pU−1(0) is the canonical projection.[8]
    3. There exist Banach spaces B1 and B2 and continuous linear maps f:X→B1, n:B1→B2, and g:B2→Y such that n:B1→B2 is nuclear and N=g∘n∘f.[10]
    4. There exists an equicontinuous sequence (xi′)i=1∞ in X′, a bounded Banach disk B⊆Y, a sequence (yi)i=1∞ in B, and a complex sequence (ci)i=1∞ such that ∑i=1∞|ci|<∞ and N is equal to the mapping:[10] N(x)=∑i=1∞cix'i(x)yi for all x∈X.
  • If X is barreled and Y is quasi-complete, then N is nuclear if and only if N has a representation of the form N(x)=∑i=1∞cix'i(x)yi with (xi′)i=1∞ bounded in X′, (yi)i=1∞ bounded in Y and ∑i=1∞|ci|<∞.[10]

Properties

[edit]

The following is a type of Hahn-Banach theorem for extending nuclear maps:

  • If E:X→Z is a TVS-embedding and N:X→Y is a nuclear map then there exists a nuclear map N~:Z→Y such that N~∘E=N. Furthermore, when X and Y are Banach spaces and E is an isometry then for any ϵ>0, N~ can be picked so that ‖N~‖Tr≤‖N‖Tr+ϵ.[18]
  • Suppose that E:X→Z is a TVS-embedding whose image is closed in Z and let π:Z→Z/Im⁡E be the canonical projection. Suppose all that every compact disk in Z/Im⁡E is the image under π of a bounded Banach disk in Z (this is true, for instance, if X and Z are both Fréchet spaces, or if Z is the strong dual of a Fréchet space and Im⁡E is weakly closed in Z). Then for every nuclear map N:Y→Z/Im⁡E there exists a nuclear map N~:Y→Z such that π∘N~=N.
    • Furthermore, when X and Z are Banach spaces and E is an isometry then for any ϵ>0, N~ can be picked so that ‖N~‖Tr≤‖N‖Tr+ϵ.[18]

Let X and Y be Hausdorff locally convex spaces.

  • Any nuclear map N:X→Y is compact.[4]
  • For every topology of uniform convergence on L(X;Y), the nuclear maps are contained in the closure of X′⊗Y (when X′⊗Y is viewed as a subspace of L(X;Y)).[8]

See also

[edit]

References

[edit]
  1. ↑ "Nuclear operator - Encyclopedia of Mathematics". encyclopediaofmath.org. Retrieved 2026-08-24.
  2. ↑ "Produits Tensoriels Topologiques et Espaces Nucléaires". pubs.ams.org. Retrieved 2026-08-24.
  3. ↑ Trèves 2006, p. 488.
  4. 1 2 3 Trèves 2006, p. 483.
  5. 1 2 Trèves 2006, p. 490.
  6. ↑ Schaefer & Wolff 1999, p. 92.
  7. 1 2 Schaefer & Wolff 1999, p. 93.
  8. 1 2 3 Schaefer & Wolff 1999, p. 98.
  9. 1 2 3 Trèves 2006, pp. 478–479.
  10. 1 2 3 4 5 Trèves 2006, pp. 481–483.
  11. 1 2 3 Trèves 2006, p. 484.
  12. ↑ Trèves 2006, pp. 483–484.
  13. ↑ Trèves 2006, pp. 488–492.
  14. 1 2 3 Trèves 2006, pp. 492–494.
  15. ↑ Trèves 2006, pp. 502–508.
  16. ↑ Trèves 2006, pp. 479–481.
  17. 1 2 Schaefer & Wolff 1999, p. 100.
  18. 1 2 Trèves 2006, p. 485.

Bibliography

[edit]
  • Diestel, Joe (2008). The metric theory of tensor products : Grothendieck's résumé revisited. Providence, R.I: American Mathematical Society. ISBN 978-0-8218-4440-3. OCLC 185095773.
  • Dubinsky, Ed (1979). The structure of nuclear Fréchet spaces. Berlin New York: Springer-Verlag. ISBN 3-540-09504-7. OCLC 5126156.
  • Grothendieck, Alexander (1966). Produits tensoriels topologiques et espaces nucléaires (in French). Providence: American Mathematical Society. ISBN 0-8218-1216-5. OCLC 1315788.
  • Husain, Taqdir (1978). Barrelledness in topological and ordered vector spaces. Berlin New York: Springer-Verlag. ISBN 3-540-09096-7. OCLC 4493665.
  • Khaleelulla, S. M. (1982). Counterexamples in Topological Vector Spaces. Lecture Notes in Mathematics. Vol. 936. Berlin, Heidelberg, New York: Springer-Verlag. ISBN 978-3-540-11565-6. OCLC 8588370.
  • Narici, Lawrence; Beckenstein, Edward (2011). Topological Vector Spaces. Pure and applied mathematics (Second ed.). Boca Raton, FL: CRC Press. ISBN 978-1584888666. OCLC 144216834.
  • Nlend, H (1977). Bornologies and functional analysis : introductory course on the theory of duality topology-bornology and its use in functional analysis. Amsterdam New York New York: North-Holland Pub. Co. Sole distributors for the U.S.A. and Canada, Elsevier-North Holland. ISBN 0-7204-0712-5. OCLC 2798822.
  • Nlend, H (1981). Nuclear and conuclear spaces : introductory courses on nuclear and conuclear spaces in the light of the duality. Amsterdam New York New York, N.Y: North-Holland Pub. Co. Sole distributors for the U.S.A. and Canada, Elsevier North-Holland. ISBN 0-444-86207-2. OCLC 7553061.
  • Pietsch, Albrecht (1972). Nuclear locally convex spaces. Berlin, New York: Springer-Verlag. ISBN 0-387-05644-0. OCLC 539541.
  • Robertson, A. P. (1973). Topological vector spaces. Cambridge England: University Press. ISBN 0-521-29882-2. OCLC 589250.
  • Ryan, Raymond (2002). Introduction to tensor products of Banach spaces. London New York: Springer. ISBN 1-85233-437-1. OCLC 48092184.
  • Schaefer, Helmut H.; Wolff, Manfred P. (1999). Topological Vector Spaces. GTM. Vol. 8 (Second ed.). New York, NY: Springer New York Imprint Springer. ISBN 978-1-4612-7155-0. OCLC 840278135.
  • Trèves, François (2006) [1967]. Topological Vector Spaces, Distributions and Kernels. Mineola, N.Y.: Dover Publications. ISBN 978-0-486-45352-1. OCLC 853623322.
  • Wong (1979). Schwartz spaces, nuclear spaces, and tensor products. Berlin New York: Springer-Verlag. ISBN 3-540-09513-6. OCLC 5126158.
[edit]