Scinovex

Research article Β· 2020 Β· Full text

International Journal of Statistics and Applied Mathematics

A note on -Isometries and self-similarity of operators on Hilbert spaces

Bernard Mutuku NzimbiStephen Wanyonyi Luketero

Volume 5(2) Β· Pages 38–42

Research areasAdvanced Topics in AlgebraHolomorphic and Operator TheoryAdvanced Differential Geometry Research

↓ Read full-text PDF0 citations

Abstract

Author-written

In this paper we investigate -isometries and related classes of operators. We will also introduce and study the notions of similitudes, self-similarity and -metric equivalence relations of operators. It will be shown that self-similarity implies -metric equivalence of operators. We will characterize -isometric and -unitary and prove that quasisimilar -isometries are unitarily equivalent.

Summary

Source-checked

Plain-language reading of the full paper, prepared with AI and checked against the text. Each point links to the authors' own words.

  1. Why it matters

    • Similitudes are geometric transformations composed of rotation, translation, and scaling that preserve the shape of sets up to a scaling factor. View evidence β†—

      Why it matters

      Similitudes are geometric transformations composed of rotation, translation, and scaling that preserve the shape of sets up to a scaling factor.

      In the paper's own words

      β€œSimilitudes are obtained as compositions of rotation, translation and scaling (dilation) operators. Such operators are similarities in the geometric sense, in that they map sets to similar sets and so preserve shapes up to a scaling factor.”

      Page 1

      Found in: Introduction

      Read it in the full textOpen PDF at page 1 β†—
    • Because similitudes alter the distance between points, they have applications in image compression (wavelets), DNA sequence analysis, gene mutation, and the study of frames. View evidence β†—

      Why it matters

      Because similitudes alter the distance between points, they have applications in image compression (wavelets), DNA sequence analysis, gene mutation, and the study of frames.

      In the paper's own words

      β€œSimilitudes in general reduce or increase distance between points and are very applicable in the study and application of frames, specifically generalized or tight π‘”βˆ’frames ([1], Definition 3.1) and wavelets in image compression and decompression schemes, geometry and DNA sequences, gene mutation and fractals (see [8]).”

      Page 5

      Found in: Discussion

      Read it in the full textOpen PDF at page 5 β†—
  2. The question

    • The authors aim to investigate alpha-isometries and introduce and study the concepts of similitudes, self-similarity, and alpha-metric equivalence relations of operators. View evidence β†—

      The question

      The authors aim to investigate alpha-isometries and introduce and study the concepts of similitudes, self-similarity, and alpha-metric equivalence relations of operators.

      In the paper's own words

      β€œIn this paper we investigate 𝛼 -isometries and related classes of operators. We will also introduce and study the notions of similitudes, self-similarity and 𝛼 -metric equivalence relations of operators.”

      Page 1

      Found in: Conclusion

      Read it in the full textOpen PDF at page 1 β†—
  3. What they did

    • The study is conducted within the framework of a Hilbert space H and the Banach algebra of bounded linear operators B(H). View evidence β†—

      What they did

      The study is conducted within the framework of a Hilbert space H and the Banach algebra of bounded linear operators B(H).

      In the paper's own words

      β€œLet 𝐻 denote a Hilbert space and 𝐡(𝐻) denote the Banach algebra of bounded linear operators.”

      Page 1

      Found in: Introduction

      Read it in the full textOpen PDF at page 1 β†—
    • The authors analyze operators by representing them as products of scalars and isometries. View evidence β†—

      What they did

      The authors analyze operators by representing them as products of scalars and isometries.

      In the paper's own words

      β€œSuppose 𝐴, 𝐡∈𝐡(𝐻) such that 𝐴= 𝛼𝑉 and𝐡= π›½π‘Š, where 𝛼, 𝛽> 0 and 𝑉, π‘Šare isometries.”

      Page 2

      Found in: T T

      Read it in the full textOpen PDF at page 2 β†—
    • The proofs rely on established properties of isometries, such as the fact that their product is also an isometry. View evidence β†—

      What they did

      The proofs rely on established properties of isometries, such as the fact that their product is also an isometry.

      In the paper's own words

      β€œThe rest of the proof follows from the fact that the product of isometries is an isometry.”

      Page 2

      Found in: T T

      Read it in the full textOpen PDF at page 2 β†—
  4. What they found

    • The authors prove that the product of two similitudes is itself a similitude, meaning the set of similitudes forms a semigroup. View evidence β†—

      What they found

      The authors prove that the product of two similitudes is itself a similitude, meaning the set of similitudes forms a semigroup.

      In the paper's own words

      β€œTheorem 2.1 The product of similitudes is a similitude.”

      Page 1

      Found in: T T

      Read it in the full textOpen PDF at page 1 β†—
    • The study finds that if two similitude operators are quasisimilar, they must also be similar. View evidence β†—

      What they found

      The study finds that if two similitude operators are quasisimilar, they must also be similar.

      In the paper's own words

      β€œTheorem 2.2 If 𝐴, 𝐡∈𝐡(𝐻) are quasisimilar similitude operators, then they are similar.”

      Page 2

      Found in: T T

      Read it in the full textOpen PDF at page 2 β†—
    • Self-similarity is established as an equivalence relation specifically within the subalgebra of invertible operators. View evidence β†—

      What they found

      Self-similarity is established as an equivalence relation specifically within the subalgebra of invertible operators.

      In the paper's own words

      β€œTheorem 2.14 Self-similarity is an equivalence relation in the Banach subalgebra of invertible operators.”

      Page 3

      Found in: Theorem 2.11

      Read it in the full textOpen PDF at page 3 β†—
    • The authors demonstrate that operators that are self-similar are also alpha-metrically equivalent. View evidence β†—

      What they found

      The authors demonstrate that operators that are self-similar are also alpha-metrically equivalent.

      In the paper's own words

      β€œTheorem 3.4 Self-similar operators are π›Όβˆ’metrically equivalent.”

      Page 3

      Found in: πΎπ‘’π‘Ÿ(𝐡) β‰  πΎπ‘’π‘Ÿ(𝐴).

      Read it in the full textOpen PDF at page 3 β†—
    • The paper proves that alpha-isometries that are quasisimilar are necessarily unitarily equivalent. View evidence β†—

      What they found

      The paper proves that alpha-isometries that are quasisimilar are necessarily unitarily equivalent.

      In the paper's own words

      β€œTheorem 3.12 Quasisimilar π›Όβˆ’isometries are unitarily equivalent.”

      Page 4

      Found in: T T

      Read it in the full textOpen PDF at page 4 β†—
    • The authors establish that every alpha-isometry is a quasinormal operator. View evidence β†—

      What they found

      The authors establish that every alpha-isometry is a quasinormal operator.

      In the paper's own words

      β€œTheorem 3.15 Everyπ›Όβˆ’ isometry 𝑇is quasinormal.”

      Page 4

      Found in: T T

      Read it in the full textOpen PDF at page 4 β†—
    • The study finds that the commutant of an alpha-unitary operator is composed of isometries. View evidence β†—

      What they found

      The study finds that the commutant of an alpha-unitary operator is composed of isometries.

      In the paper's own words

      β€œCorollary 4.2 The commutant of an π›Όβˆ’unitary operator 𝑇 consists of isometries.”

      Page 4

      Found in: Conclusion

      Read it in the full textOpen PDF at page 4 β†—
    • In a geometric context, the authors note that for figures similar with a scale factor alpha, the perimeter ratio is alpha and the area ratio is alpha squared. View evidence β†—

      What they found

      In a geometric context, the authors note that for figures similar with a scale factor alpha, the perimeter ratio is alpha and the area ratio is alpha squared.

      In the paper's own words

      β€œProposition 5.4 If two figures are similar with a scale factor𝛼, then the perimeter ratio is 𝛼and the area ratio is 𝛼2.”

      Page 4

      Found in: πΎπ‘’π‘Ÿ(𝐡) β‰  πΎπ‘’π‘Ÿ(𝐴).

      Read it in the full textOpen PDF at page 4 β†—
  5. What it means

    • The authors conclude that they have successfully introduced alpha-isometry and self-similarity, proving that the latter is an equivalence relation and that quasisimilar alpha-isometries are unitarily equivalent. View evidence β†—

      What it means

      The authors conclude that they have successfully introduced alpha-isometry and self-similarity, proving that the latter is an equivalence relation and that quasisimilar alpha-isometries are unitarily equivalent.

      In the paper's own words

      β€œIn this paper, the notions of π›Όβˆ’isometry and self-similarity are introduced. It has been shown that self-similarity is an equivalence relation. It has also been shown that quasisimilar π›Όβˆ’isometries are unitarily equivalent. It is also shown that the commutant of an π›Όβˆ’ unitary operator T consists of isometries.”

      Page 5

      Found in: Conclusion

      Read it in the full textOpen PDF at page 5 β†—

Visual summary

1 visuals

Illustrations explain the paper; they do not show study data.

AI-generated illustration
AI-generated conceptual illustration (Two identical geometric polygons of different sizes in a 3D coordinate space). It explains the idea and does not show study data.β†— Enlarge
Geometric Similitude
AI-generated conceptual illustration (Two identical geometric polygons of different sizes in a 3D coordinate space). It explains the idea and does not show study data.

AI-generated conceptual illustration (Two identical geometric polygons of different sizes in a 3D coordinate space). It explains the idea and does not show study data.

Open full-size image β†—

Geometric Similitude

AI-generated conceptual illustration (Two identical geometric polygons of different sizes in a 3D coordinate space). It explains the idea and does not show study data.

View source evidence β†—
FULL TEXT Original article Β· reflowed for reading

Introduction

Let 𝐻 denote a Hilbert space and 𝐡(𝐻) denote the Banach algebra of bounded linear operators. If 𝑇 ∈ 𝐡(𝐻), then 𝑇 * denotes the adjoint of 𝑇 while πΎπ‘’π‘Ÿ(𝑇), π‘…π‘Žπ‘›(𝑇), 𝑀and 𝑀 βŠ₯ stands for the kernel of𝑇, range of , closure of 𝑀 and orthogonal complement of a closed subspace 𝑀 of𝐻, respectively. We denote by 𝜎(𝑇), ‖𝑇‖, π‘Ÿ(𝑇) and π‘Š(𝑇) the spectrum, norm, spectral radius of and numerical range of 𝑇, respectively. Recall that an operator 𝑇 ∈ 𝐡(𝐻) is normal if 𝑇 * 𝑇 = 𝑇𝑇 * , unitary if 𝑇 * 𝑇 = 𝑇𝑇 * = 𝐼, a projection(or idempotent) if 𝑇 2 = 𝑇, an orthogonal projection if 𝑇 2 = 𝑇and 𝑇 * = 𝑇, an isometry if 𝑇 * 𝑇 = 𝐼 (equivalently‖𝑇π‘₯ -𝑇𝑦‖ = β€–π‘₯ -𝑦‖, for all π‘₯, 𝑦 ∈ 𝐻), a partial isometry if 𝑇𝑇 * 𝑇 = 𝑇, quasinormal if (𝑇 * 𝑇)𝑇 = 𝑇(𝑇 * 𝑇), a scalar if 𝑇 = 𝛼𝐼 where 𝛼 ∈ β„‚, a contraction if ‖𝑇π‘₯β€– ≀ β€–π‘₯β€–, for allπ‘₯ ∈ 𝐻. Two operators 𝐴 ∈ 𝐡(𝐻) and 𝐡 ∈ 𝐡(𝐾) are said to be similar if there exists an invertible operator 𝑁 ∈ 𝐡(𝐻, 𝐾) such that 𝑁𝐴 = 𝐡𝑁or equivalently𝐴 = 𝑁 -1 𝐡𝑁, and are unitarily equivalent if there exists a unitary operator π‘ˆ ∈ 𝐡 + (𝐻, 𝐾) (Banach algebra of all invertible operators in𝐡(𝐻)) such that π‘ˆπ΄ = π΅π‘ˆ (i.e.𝐴 = π‘ˆ * π΅π‘ˆequivalently, 𝐴 = π‘ˆ -1 π΅π‘ˆ). An operator 𝑋 ∈ 𝐡(𝐻, 𝐾) is a quasiaffinity if it is injective with dense range. Two operators 𝐴 ∈ 𝐡(𝐻) and 𝐡 ∈ 𝐡(𝐾) are quasisimilar if there exist quasiaffinities 𝑋 ∈ 𝐡(𝐻, 𝐾) and π‘Œ ∈ 𝐡(𝐾, 𝐻) such that 𝑋𝐴 = 𝐡𝑋 and π΄π‘Œ = π‘Œπ΅. Two operators 𝐴, 𝐡 ∈ 𝐡(𝐻) are said to be metrically equivalent if‖𝐴π‘₯β€– = ‖𝐡π‘₯β€–, (equivalently, |⟨𝐴π‘₯, 𝐴π‘₯⟩| 1 2 = |⟨𝐡π‘₯, 𝐡π‘₯⟩| 1 2 for allπ‘₯ ∈ 𝐻, which is equivalent to having 𝐴 * 𝐴 = 𝐡 * 𝐡). The notion of metric equivalence of operators was introduced by Nzimbi et al. [6] . Clearly similarity, unitary equivalence and metric equivalence are equivalence relations on𝐡(𝐻). A similitude is an operator 𝑇 ∈ 𝐡(𝐻) that is a positive scalar multiple of an isometry (equivalently, if ‖𝑇π‘₯ -𝑇𝑦‖ = 𝛼‖π‘₯ -𝑦‖, for all π‘₯, 𝑦 ∈ 𝐻 and 𝛼 > 0). Clearly every isometry is a similitude. Similitudes are obtained as compositions of rotation, translation and scaling (dilation) operators. Such operators are similarities in the geometric sense, in that they map sets to similar sets and so preserve shapes up to a scaling factor. Similitudes are extensions of isometries. The theory of isometries has been extensively studied (see [2] , [4] and [5] for more exposition).

Main results Theorem 2.1

The product of similitudes is a similitude.

T T

Proof. Suppose 𝐴, 𝐡 ∈ 𝐡(𝐻) such that 𝐴 = 𝛼𝑉 and𝐡 = π›½π‘Š, where 𝛼, 𝛽 > 0 and 𝑉, π‘Šare isometries. Then 𝐴𝐡 = π›Όπ›½π‘‰π‘Š. The rest of the proof follows from the fact that the product of isometries is an isometry. Theorem 2.1 says that the set of similitudes is a semigroup. Theorem 2.2 If 𝐴, 𝐡 ∈ 𝐡(𝐻) are quasisimilar similitude operators, then they are similar. Proof. Suppose, without loss of generality, that 𝐴 = π›Όπ‘ˆand 𝐡 = 𝛼𝑉, where π‘ˆ, 𝑉 are isometries and 𝛼 > 0. Using the fact that every isometry is quasinormal and ( [9] , Proposition 4.2), we conclude that any quasisimilar isometries in a Hilbert space are unitarily equivalent. Supposeπ‘ˆ = π‘Šπ‘‰π‘Š * , where π‘Šis unitary. Then𝐴 = π›Όπ‘ˆ = π›Όπ‘Š( 1 𝛼 𝐡)π‘Š * = π‘Šπ΅π‘Š * = π‘Šπ΅π‘Š -1 . This proves the claim. A consequence of Theorem 2.2 is that quasisimilar similitudes are unitarily equivalent. Definition 2.3 Two operators 𝐴, 𝐡 ∈ 𝐡(𝐻) are said to be self-similar denoted by 𝐴 ∼ 𝑠.𝑠 𝐡 if there is a similitude 𝑆 ∈ 𝐡(𝐻) such that𝐡 = 𝑆𝐴. A closed subspace βŠ† 𝐻is said to be self-similar if it is invariant under a similitude𝑇. Remark. We note that the trivial subspaces 𝑀 1 = {0}and 𝑀 2 = 𝐻are invariant under any bounded linear operator on 𝐻and hence are self-similar.

In particular, ifπœ€ = 0, then the class of 0 -isometries contains the class of isometries and contraction operators. A similitude need not be invertible, but if it is invertible, then it is a multiple if a unitary operator. Theorem 2.8 [5] An operator 𝑇 is a multiple of an isometry if and only if 𝑇 * 𝑇 is a scalar. Theorem 2.9 If 𝑇 is a similitude, then 𝜎(|𝑇|) = {𝛼} for some 𝛼 > 0. Proof. This follows from the fact that 𝑇 is a similitude if and only if |𝑇| = 𝛼𝐼, for some 𝛼 > 0. Definition 2.10 An operator 𝑇 ∈ 𝐡(𝐻) is called a 𝛼 -projection if 𝑇 = 𝛼𝑃, for some projection 𝑃 ∈ 𝐡(𝐻). An operator 𝑇 ∈ 𝐡(𝐻) is called a 𝛼 -isometry if 𝑇 = 𝛼𝑉, where 𝑉is an isometry and 𝛼 ∈ β„‚. It is clear that if 𝐴 ∈ 𝐡(𝐻) is an 𝛼 -isometry and π‘ˆis an isometry, then 𝐡 = π‘ˆπ΄π‘ˆ * is a 𝛼 -projection. It is also clear that an 𝛼 -projection is a projection if 𝛼 ∈ {0,1}. It is also clear that a similitude is an 𝛼 -isometry. In fact the following inclusion holds: π‘†π‘–π‘šπ‘–π‘™π‘–π‘‘π‘’π‘‘π‘’ βŠ‚ πœ€ -π‘–π‘ π‘œπ‘šπ‘’π‘‘π‘Ÿπ‘¦ βŠ‚ 𝛼 -π‘–π‘ π‘œπ‘šπ‘’π‘‘π‘Ÿπ‘¦.

Theorem 2.11

If 𝐴 ∈ 𝐡(𝐻) is unitarily equivalent to a similitude, then it is a similitude. Proof. Suppose that 𝐴 = π‘ˆπ΅π‘ˆ * , where 𝐡 = 𝛼𝑉 with 𝑉 isometric. Then a simple computation shows that 𝐴 = π‘ˆπ΅π‘ˆ * = π‘ˆ(𝛼𝑉)π‘ˆ * = π›Όπ‘ˆπ‘‰π‘ˆ * . Now, if we let π‘Š = π‘ˆπ‘‰π‘ˆ * , it is clear that π‘Šis an isometry. Therefore 𝐴 = π›Όπ‘Š. This establishes the claim.

Example 1 For 𝛼 > 0, the operator 𝑇acting on β„“ 2 (β„•) defined by 𝑇(π‘₯ 1 , π‘₯ 2 , π‘₯ 3 , β‹― ) = (0, 𝛼π‘₯ 1 , 𝛼π‘₯ 2 , 𝛼π‘₯ 3 , β‹― ) is a similitude. Clearly, 𝑇 = 𝛼𝑉, where 𝑉is the unilateral shift on β„“ 2 (β„•). In fact the set of similitudes is a multiplicative group.

Theorem 2.14 Self-similarity is an equivalence relation in the Banach subalgebra of invertible operators. Proof. Recall that two operators 𝐴, 𝐡 ∈ 𝐡(𝐻) are self-similar if there exists a similitude 𝑆 ∈ 𝐡(𝐻) such that 𝐴 = 𝑆𝐡. Clearly, 𝐴 = 𝐼𝐴. Thus 𝐴 ∼ 𝑠.𝑠 𝐴. Suppose 𝐴, 𝐡 ∈ 𝐡(𝐻) are invertible and 𝐴 ∼ 𝑠.𝑠 𝐡. Then there exists an invertible similitude𝑆 ∈ 𝐡(𝐻)

such that 𝐴 = 𝑆𝐡. Using Proposition 2.13, we have that 𝐡 = 𝑆 -1 𝐴. This shows that 𝐡 ∼

πΎπ‘’π‘Ÿ(𝐡) β‰  πΎπ‘’π‘Ÿ(𝐴).

A simple computation shows that these operators are similar but not self-similar. This shows that there exist similar operators which are not self-similar. We also note that while similarity preserves invertibility, self-similarity need not preserve invertibility. To see this, let 𝐴 = 𝐼and 𝐡 = 𝑆, the unilateral shift on β„“ 2 (β„•). Clearly,𝐴 and 𝐡are self-similar but are not both invertible.

Theorem 3.3 If 𝐴, 𝐡 ∈ 𝐡(𝐻) are similar and self-similar, then 𝐴 = 𝐡.

Theorem 3.4 Self-similar operators are 𝛼 -metrically equivalent.

Proof. Suppose that 𝐴, 𝐡 ∈ 𝐡(𝐻) are such that 𝐴 = 𝑆𝐡for some similitude𝑆 = 𝛼𝑉, where 𝑉is an isometry and 𝛼 > 0. A simple computation gives 𝐴 * 𝐴 = 𝛼 2 𝐡 * 𝐡, which establishes the claim. We note that if 𝑇 ∈ 𝐡(𝐻, 𝐾) is an 𝛼 -isometry, then βŸ¨π‘‡π‘₯, π‘‡π‘¦βŸ© = βŸ¨π›Όπ‘₯, π›Όπ‘¦βŸ© = |𝛼| 2 ⟨π‘₯, π‘¦βŸ©, for all π‘₯ ∈ 𝐻and𝑦 ∈ 𝐾. In this case, we say that 𝑇is 𝛼 -isometric. Clearly, every unitary operator is an 𝛼 -unitary and every 𝛼 -unitary is an 𝛼 -isometry. That is

The converse is not true in general. The unilateral shift on β„“ 2 (β„•) is an 𝛼 -isometry which is not an 𝛼 -unitary. Example 4. Suppose that in the Hilbert space 𝐻 = ℝ 2 , two closed and connected subspaces 𝛺 1 and 𝛺 2 are similar, with a scaling factor 𝛼 > 0. Then there exists a similitude𝑇: 𝛺 1 β†’ 𝛺 2 . A simple computation shows that the perimeter ratio is 𝛼and the area ratio is 𝛼 2 .

Remark. Example 4 can be extended to𝛼 -isometries (that is, all𝛼 ∈ β„‚). In any case, the distance between points is |𝛼|. This is the ratio of magnification or dilation factor or similitude factor.

Discussion

Similitudes that are contractions (that is, 𝜢 -isometries where the scaling factor0 < 𝛼 < 1 ) are applicable in the study of fractals and are a popular tool to provide existence of solutions in many problems. Self-similarity is a characterization of fractals arising from iterated function systems and find application in chaotic dynamical systems and modelling of systems (see [10] ). Similitudes in general reduce or increase distance between points and are very applicable in the study and application of frames, specifically generalized or tight 𝑔 -frames ( [1] , Definition 3.1) and wavelets in image compression and decompression schemes, geometry and DNA sequences, gene mutation and fractals (see [8] ).

Conclusion

In this paper, the notions of 𝛼 -isometry and self-similarity are introduced. It has been shown that self-similarity is an equivalence relation. It has also been shown that quasisimilar 𝛼 -isometries are unitarily equivalent. It is also shown that the commutant of an 𝛼unitary operator T consists of isometries.

Acknowledgements

The author wishes to pay tribute to the Eastern Africa University Mathematics Programme (EAUMP) for funding this research.

References

10 in the original paper
  1. Avazzadeh SH, Kamyabi-Gol RA, Raisi Tousi R. Continuous frames and g-frames, Bull. Iranian Math. Soc. 2014; 40(4):1047-1055.
  2. Paul S Bourdon, Similarity of parts to the whole for certain multiplication operators, Proc. Amer. Math. Soc. 1987; 99(3):563-567.
  3. Hoover TB. Quasisimilarity of operators, Ill. J of Math. 1972; 16:678-686.
  4. Kubrusly CS. An introduction to models and decompositions in operator theory, Birkhauser, Boston, 1997.
  5. Kubrusly CS. Hilbert space operators: A problem solving approach, Birkhauser, Basel, Boston, 2003.
  6. Nzimbi BM, Pokhariyal GP, Moindi SK. A note on metric equivalence of some operators, Far East J of Math. Sci. (FJMS). 2013; 75(2):301-318.
  7. Radjabalipour M, Rai SN. On commutators of isometries and hyponormal operators, J Sci. Islamic. Republic of. Iran. 1989; 1(1):60-62.
  8. Strichartz RS. Fractals in the large, Canadian J Math. 1998; 50:638-657.
  9. Szπ‘’Μˆcs E. Similitude and Modellinh, Fundamental Studies in Engineering 2, Akad𝑒 β€² miai Kiadπ‘œ β€² , Budapest, 1980.
  10. Wu PY. Hyponormal operators quasisimilar to an isometry, Trans. Amer. Math. Soc. 1985; 291:229-239.