Front. Appl. Math. Stat. Frontiers in Applied Mathematics and Statistics Front. Appl. Math. Stat. 2297-4687 Frontiers Media S.A. 10.3389/fams.2024.1397374 Applied Mathematics and Statistics Original Research Darboux transformation of symmetric Jacobi matrices and Toda lattices Kovalyov Ivan 1 * Levina Oleksandra 2 1Institut für Mathematik, Universität Osnabrück, Osnabrück, Germany 2Faculty of Mathematics, Informatics and Physics, Mykhailo Drahomanov Ukrainian State University, Kyiv, Ukraine

Edited by: Marina Chugunova, Claremont Graduate University, United States

Reviewed by: Youssri Hassan Youssri, Cairo University, Egypt

Dudkin Mykola, Kyiv Polytechnic Institute, Ukraine

 *Correspondence: Ivan Kovalyov i.m.kovalyov@gmail.com 16 05 2024 2024 10 1397374 07 03 2024 25 04 2024 Copyright © 2024 Kovalyov and Levina. 2024 Kovalyov and Levina

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Let J be a symmetric Jacobi matrix associated with some Toda lattice. We find conditions for Jacobi matrix J to admit factorization J = LU (or J = 𝔘𝔏) with L (or 𝔏) and U (or 𝔘) being lower and upper triangular two-diagonal matrices, respectively. In this case, the Darboux transformation of J is the symmetric Jacobi matrix J(p) = UL (or J(d) = 𝔏𝔘), which is associated with another Toda lattice. In addition, we found explicit transformation formulas for orthogonal polynomials, m-functions and Toda lattices associated with the Jacobi matrices and their Darboux transformations.

 Jacobi matrix Darboux transformation orthogonal polynomials moment problem Toda lattice section-at-acceptance Numerical Analysis and Scientific Computation 1 Introduction

Let a sequence of real numbers s={sn}n=0∞ be associated with a measure μ on (−∞, +∞), i.e.

sn=∫-∞+∞λndμ(λ),  n∈ℤ+.

However, in the general case, s={sn}n=0∞ is associated with a linear functional 𝔖 by

sn=𝔖(λn),  n∈ℤ+.

We consider the sequence s={sn}n=0∞ such that

Dn≠0,   for all  n∈ℕ,

where Dn=det(si+j)i,j=0n-1. Note, if Dn > 0 for all n ∈ ℕ, then there exists measure μ associated with s={sn}n=0∞, otherwise, the sequence s={sn}n=0∞ is associated with only linear functional 𝔖.

On the other hand (see [1, 2]), the real sequence s={sn}n=0∞ is associated with the symmetric Jacobi matrix J and the sequence of orthogonal polynomials of the first kind {Pn(λ)}n=0∞, which can defined by

P0(λ)≡1  and  Pn(λ)=1Dn-1Dn|s0s1…sns1s2…sn+1…………sn-1sn…s2n-11λ…λn|.

[3, 4] Moreover, the sequence {Pn(λ)}n=0∞ satisfies a three-term recurrence relation

λPn(λ)=an+1Pn+1(λ)+bnPn(λ)+anPn-1(λ)

with the initial conditions

P-1(λ)≡0  and  P0(λ)≡1.

In the short form we can rewrite Equation (1.2) as

JP(λ)=λP(λ),

where P(λ)=(P0(λ),…,Pn(λ),…)T and the symmetric Jacobi matrix J is defined by

J=(b0a1a1b1a2a2b2⋱⋱⋱).

On the other hand, the symmetric Jacobi matrix J is associated with the moment sequence s={sn}n=0∞, the following relation holds (see [2, 5])

sn=(e0,Jne0)   for all  n∈ℤ+,

where e0=(1,0,…)T and m– function of Jacobi matrix is found by

m(z)=∫ℝdμ(λ)λ-z.

There exist two type transformations of orthogonal polynomials, which are the Christoffel and Geronimus transformations. One are studied in the paper Zhedanov [6]. The Christoffel transformation is defined by

P~(λ)=Pn+1(λ)-AnPn(λ)λ-α,  n∈ℤ+,

where An=Pn+1(α)Pn(α) and α is arbitrary parameter. Moreover, Equation (1.7) can be rewritten as follows:

Theorem 1.1. ([7, Theorem 1.5]) Let {Pn(λ)}n=0∞ be the sequence of the orthogonal polynomials associated with Equation (1.2). Then the Christoffel–Darboux formula takes the following form

∑i=0nPi(x)Pi(t)=an+1Pn+1(x)Pn(t)-Pn(x)Pn+1(t)x-t.

The second transformation is a Geronimus transformation of the orthogonal polynomials [6], one is defined by

P~(λ)=Pn(λ)-BnPn-1(λ),  Bn∈ℝ  and  n∈ℕ.

Toda lattice. The Toda lattice is a system of differential equations

xn″(t)=exn-1-xn-exn-xn+1,  n∈ℕ,

which was introduced in Toda [8].

We study the semi-infinite system with x−1 = −∞. [9, 10] Flaschka variables are defined by

ak=12exk-1-xk2  and  bk=-12xk′.

Therefore, we obtain the following system in terms of Flaschka variables

ak′=ak(bk-bk-1)  and  bk′=2(ak+12-ak2),  a0=0.

Hence, the semi-infinite Toda lattice is associated with the symmetric Jacobi matrix J and Lax pair (J, A), such that

[J,A]=JA-AJ,

where the matrix A = J+ − J−, where J+ and J− are upper and lower triangular part of J, respectively and

A=(0a1-a10a2-a20⋱⋱⋱).

As is known (see [8, 11]), the system (1.11) is equivalent to the following

J′=-[J,A].

Darboux transformation of the monic classical and generalized Jacobi matrices were studied in Bueno and Marcellán [12], Derevyagin and Derkach [13], and Kovalyov [14, 15]. Darboux transformation involves finding a factorization of a matrix from a certain class such that the new matrix is from the same class. There are two types of Darboux transformation: transformation with and without parameter. Jacobi matrix is associated with many objects. There are moment sequence, measure, linear functional orthogonal polynomials and Toda lattice. Hence, in the current paper, we study not only Darboux transformation of the symmetric Jacobi matrices, but we also study the transformation of the associated objects. Hence, we investigate the Darboux transformation of the symmetric Jacobi matrices J and find relations between associated Toda lattice, orthogonal polynomials, moment sequences and m–functions. We obtain that the Darboux transformation without parameter of the symmetric Jacobi matrices has more additional existence conditions in contrast to case of the monic Jacobi matrices. On the other hand, the Darboux transformation with parameter of the symmetric Jacobi matrices is generated more easily. The results obtained can be applied for further research related to symmetric Jacobi matrices, Toda lattices and inverse problems. Of course, it can also be applied to the Toda lattice hierarchy.

Now, briefly describe the content of the paper. Section 2 contains Darboux transformation without parameter of the symmetric Jacobi matrix J. We find LU–factorization of J and the transformed matrix J(p). Relation between Toda lattices, moment sequences and m–functions associated with the Jacobi matrices was obtained. In this case, the orthogonal polynomials are transformed by the Christoffel formula (1.7). In Section 3, we study the Darboux transformation with parameter of the symmetric Jacobi matrix J. We find 𝔘𝔏–factorization of J and transformed matrix J(d). Moreover, the relations between orthogonal polynomials, m–functions, moment sequence and Toda lattices are found according to explicit formulas.

 2 Darboux transformation without parameter of symmetric Jacobi matrix

Now we study a Darboux transformation without parameter of symmetric Jacobi matrix J. The goal is to find the transformations of polynomials of the first kind, m-functions, measure, moment sequence and Toda lattice, which are associated with the transformed Jacobi matrix.

2.1 <italic>LU</italic>–factorization

Lemma 2.1. Let J be a symmetric Jacobi matrix. Then J admits LU–factorization

J=LU,

where L and U are lower and upper triangular matrices, respectively, which are defined by

L=(1l11l21⋱⋱)  and  U=(u1v1u2v2u3⋱⋱),

if and only if the following system is solvable

b0=u1,  v1=a1,  vj=aj,  ljuj=aj,ljvj+uj+1=bj,  uj≠0  and  lj≠0,  j∈ℕ.

Proof. Let us calculate the product LU

LU=(u1v1l1u1l1v1+u2v2l2u2l2v2+u3⋱⋱⋱).

Comparing the product LU with the Jacobi matrix J

(b0a1a1b1a2a2b2⋱⋱⋱)=(u1v1l1u1l1v1+u2v2l2u2l2v2+u3⋱⋱⋱),

we obtain the system (2.3).

If the system (2.3) is solvable, then J admits the factorization J = LU of the form (2.1–2.3), where L and U are found uniquely. Conversely, if J admit LU—factorization then the system (2.3) is solvable. This completes the proof.

Lemma 2.2. Let J be the symmetric Jacobi matrix and let J = LU be its LU– factorization of the form (2.1–2.3). Let Pj be the polynomials of the first kind associated with the matrix J. Then

Pn(0)Pn-1(0)=-1ln,   n∈ℕ.

Proof. Let J admit the LU–factorization of the form (2.1–2.3). Setting λ = 0 in Equation (1.2), we obtain

an+1Pn+1(0)+bnPn(0)+anPn-1(0)=0.

By induction, we prove Equation (2.4).

1. Let n = 0, then

a1P1(0)+b0P0(0)+a0P-1(0)=0

and due to the initial condition (1.3) and (2.3), we get

a1P1(0)+b0P0(0)=0⇒P1(0)P0(0)=-b0a1=-u1l1u1=-1l1.

2. Let n = 1, then

a2P2(0)+b1P1(0)+a1P0(0)=0

and by Equation (2.3), we have

P2(0)P1(0)+b1a2+a1a2P0(0)P1(0)=0⇒P2(0)P1(0)   =-b1a2+a1l1a2=-l12u1-u2+l12u1l2u2=-1l2.

3. Let Equation (2.4) hold for n = k − 1.

4. Let us prove Equation (2.4) for n = k, we obtain

ak+1Pk+1(0)+bnPk(0)+akPk-1(0)=0. Pk+1(0)Pk(0)+bkak+1+akak+1·Pk-1(0)Pk(0)=0.

Consequently

Pk+1(0)Pk(0)=-bkak+1-akak+1·Pk-1(0)Pk(0)

= {by Section (2.3)}

=-bk+aklkak+1=-lk2uk-uk+1+lk2uklk+1uk+1=-1lk+1.

So, Equation (2.4) is proven. This completes the proof.

Corollary 2.3. Let J be the symmetric Jacobi matrix and let J = LU be its LU–factorization of the form (2.1–2.3). Let Pj be the polynomials of the first kind associated with the matrix J. Then

Pn(0)=(-1)n∏i=1n1li.

Proof. Let J admit the LU–factorization of the form (2.1–2.3) and let Pj be the polynomials of the first kind associated with J. By Lemma 2.2, Equation (2.4) holds and we obtain

Pn(0)=Pn(0)Pn-1(0)·Pn-1(0)Pn-2(0)·…·P1(0)P0(0)=(-1)n∏i=1n1li.

So, Equation (2.5) is proven. This completes the proof.

Corollary 2.4. Let J be the symmetric Jacobi matrix and let J = LU be its LU–factorization of the form (2.1–2.3). Let Pj be the polynomials of the first kind associated with the matrix J. Then

Pn(0)=(-1)k1ln·1ln-1·…·1ln-(k-1)Pn-k(0).

Proof. Let J admit the LU–factorization of the form (2.1–2.3). By Lemma 2.2, we obtain

Pn(0)=Pn(0)Pn-1(0)·Pn-1(0)Pn-2(0)·…·Pn-k-1(0)Pn-k(0)·Pn-k(0)=(-1)k1ln·1ln-1·…·1ln-(k-1)Pn-k(0).

Hence, Equation (2.6) is proven. This completes the proof.

Theorem 2.5. Let J be the symmetric Jacobi matrix and let Pj be the polynomials of the first kind associated with J. Then J admits LU—factorization of the form (2.1–2.3) if and only if

Pj(0)≠0   for all  j∈ℤ+.

Furthermore,

b0=u1,  vj=aj,  lj=-Pj-1(0)Pj(0)  and  uj=-ajPj(0)Pj-1(0).

Proof. Let Pj(0) ≠ 0 for all j ∈ ℤ+. By Lemma 2.2 the system (2.8) is equivalent to the system (2.3). Consequently, by Lemma 2.1 the Jacobi matrix J admits LU—factorization of the form (2.1–2.3). Conversely, if the Jacobi matrix J admits LU—factorization of the form (2.1–2.3), then by Lemma 2.1 and Lemma 2.2 the polynomials of the first kind Pj satisfy (2.7). This completes the proof.

 2.2 Transformed Jacobi matrix <italic>J</italic><sup>(<italic>p</italic>)</sup> = <italic>UL</italic>

Definition 2.6. Let the symmetric Jacobi matrix J admit LU—factorization of the form (2.1–2.3). Then a transformation

J=LU→UL=J(p)

is called a Darboux transformation without parameter of the symmetric Jacobi matrix J.

Theorem 2.7. Let J be the symmetric Jacobi matrix (1.4) and let J = LU be its LU–factorization of the form (2.1–2.3). Then the Darboux transformation without parameter of the matrix J is the symmetric Jacobi matrix

J(p)=UL=(b1a1a1b2a2a2b3⋱⋱⋱)

if and only if

uj=b0  and  aj2+b02b0=bj   for all  j∈ℕ.

Proof. Calculating UL, we obtain

J(p)=UL=(u1v1u2v2u3⋱⋱)(1l11l21⋱⋱)==(u1+v1l1v1l1u2u2+v2l2v2l2u3u3+v3l3⋱⋱⋱)

= {by Equation (2.3)}

=(u1+v1l1a1l1u2u2+v2l2a2l2u3u3+v3l3⋱⋱⋱).

Consecuently, J(p) is the symmetric Jacobi matrix if and only if

ljuj+1=aj   for all  j∈ℕ.

Comparing Equation (2.3) with Equation (2.11), we get

ljuj=aj=ljuj+1⇒uj=uj+1⇒uj=b0   for all  j∈ℕ.

By Equation (2.3), uj + vjlj = bj for all j ∈ ℕ, we obtain Equations (2.9, 2.10) and J(p) is the symmetric Jacobi matrix. This completes the proof.

Theorem 2.8. Let the symmetric Jacobi matrix J satisfy (2.7) and let J = LU be its LU–factorization of the form (2.1–2.3). Let J(p) = UL be the Darboux transformation without parameter of J. Then the polynomials of the first kind Pn(p) associated with J(p) can be found by Christoffel–Darboux formula

Pn(p)(λ)=1Pn(0)Pn+1(λ)Pn(0)-Pn(λ)Pn+1(0)λ,

where Pj are the polynomials of the first kind associated with the symmetric Jacobi matrix J.

Proof. Let the Jacobi matrix J satisfy (2.7) and admit LU–factorization of the form (2.1–2.3). Calculating the inverse matrix of L, we obtain

L-1=(1-l11l1l2-l21-l1l2l3l2l3-l31⋱⋱⋱⋱⋱(-1)n∏i=1nli(-1)n-1∏i=2nli…ln-1ln-ln1⋱⋱⋱⋱⋱⋱⋱).

On the other hand,

J(p)P(λ)=ULP(p)(λ)=λP(p)(λ)⇒⇒LULP(p)(λ)=J(LP(p)(λ))=λ(LP(p)(λ))=λP(λ).

Consequently, we obtain the relation between the polynomials of the first kind

P(p)(λ)=L-1P(λ)=(1-l11l1l2-l21-l1l2l3l2l3-l31⋱⋱⋱⋱⋱)(P0(λ)P1(λ)P2(λ)P3(λ)⋮)=(P0(λ)P1(λ)-l1P0(λ)P2(λ)-l2P1(λ)+l1l2P0(λ)P3(λ)-l3P2(λ)+l2l3P1(λ)-l1l2l3P0(λ)⋮)=(P0(p)(λ)P1(p)(λ)P2(p)(λ)P3(p)(λ)⋮).

By Corollary 2.4, we obtain

Pn(p)(λ)=Pn(λ)+∑i=0n-1(-1)n-iPi(λ)∏j=i+1nlj=Pn(λ)+∑i=0n-1Pi(0)Pn(0)Pi(λ).

However, we can rewrite Equation (2.13) and by Christoffel–Darboux formula (1.8), we obtain

Pn(p)(λ)=Pn(λ)+∑i=0n-1Pi(0)Pn(0)Pi(λ)=1an+1Pn(0)∑i=0nPi(0)Pi(λ)==1Pn(0)Pn+1(λ)Pn(0)-Pn(λ)Pn+1(0)λ.

Hence, Equation (2.12) holds. This completes the proof.

In the following statements we find the connection between orthogonal polynomials, moment sequences, measures, linear functionals, m–functions and Toda lattices according to the transformation Darboux transformation without parameter of the symmetric Jacobi matrix.

Proposition 2.9. Let the symmetric Jacobi matrix J admit LU– factorization of the form (2.1–2.3) and let the symmetric Jacobi matrix J(p) = UL be the Darboux transformation without parameter of J. Let s={sn}n=0∞ and s(p)={sn(p)}n=0∞ be the moment sequences associated with the matrices J and J(p), respectively. Then the moment sequence s(p)={sn(p)}n=0∞ can be found by the following formula

sn-1(p)=snb0   for all  n∈ℕ.

Proof. Let the symmetric Jacobi matrix J admit LU–factorization of the form (2.1–2.3) and let the symmetric Jacobi matrix J(p) = UL be its Darboux transformation without parameter. By Equation (1.5), we obtain

sn=(e0,Jne0)=(e0,(LU)ne0)=(e0,L(UL)n-1Ue0)==(LTe0,(J(p))n-1b0e0)=b0(e0,(J(p))n-1e0)=b0sn-1(p).

Consequently, the moments sn-1(p) can be found by Equation (2.14). This completes the proof.

Corollary 2.10. Let the symmetric Jacobi matrix J admit LU–factorization of the form (2.1)–(2.3) and let the symmetric Jacobi matrix J(p) = UL be the Darboux transformation without parameter of J. Let 𝔖 and 𝔖(p) be the linear functionals associated with the matrices J and J(p), respectively. Then

𝔖(p)=λb0𝔖.

Proof. Let 𝔖 and 𝔖(p) be the linear functionals associated with the symmetric Jacobi matrices J = LU and J(p) = UL, respectively, where L and U are defined by Equations (2.1–2.3). By Equation (1.1), we obtain

𝔖(p)(λn-1)=sn-1(p)=snb0=1b0𝔖(λn)   for all  n∈ℕ.

Consequently, Equation (1.19) holds. This completes the proof.

Corollary 2.11. Let the symmetric Jacobi matrix J admit LU–factorization of the form (2.1–2.3) and let the symmetric Jacobi matrix J(p) = UL be the Darboux transformation without parameter of J. Let dμ and dμ(p) be the measures associated with the matrices J and J(p), respectively. Then

dμ(p)(λ)=λb0dμ(λ).

Proof. Let μ and μ(p) be the measures associated with the symmetric Jacobi matrices J = LU and J(p) = UL, respectively, where L and U are defined by Equation (2.1–2.3). Then

∫-∞+∞λn-1dμ(p)(λ)=sn-1(p)=snb0=1b0∫-∞+∞λndμ(λ)   for all  n∈ℕ.

Consequently, we find transformation of the measure and Equation (2.16) holds. This completes the proof.

Proposition 2.12. Let the symmetric Jacobi matrix J admit LU– factorization of the form (2.1–2.3) and let the symmetric Jacobi matrix J(p) = UL be the Darboux transformation without parameter of J. Let m and m(p) be the m–functions associated with the matrices J and J(p), respectively. Then

m(p)(z)=s0+zm(z)b0.

Proof. By Equation (1.6)

m(p)(z)=∫-∞+∞dμ(p)(λ)λ-z==1b0∫-∞+∞λdμ(λ)λ-z=1b0∫ℝλ-zλ-zdμ(λ)+1b0∫-∞+∞zdμ(λ)λ-z=1b0∫-∞+∞1dμ(λ)+zb0∫-∞+∞dμ(λ)λ-z=s0+zm(z)b0.

Hence, m–function is transformed by Equation (2.17). This completes the proof.

Toda latice. The last statement is the following theorem of this section. One is described the Toda lattice associated with the symmetric Jacobi matrices J(p).

Theorem 2.13. Let the symmetric Jacobi matrix J admit LU–factorization of the form (2.1–2.3) and J be associated with the Toda lattice (1.9–1.11). Let the symmetric Jacobi matrix J(p) = UL be the Darboux transformation without parameter of J. Then J(p) is associated with the following Toda lattice

xk″(t)=exk-1-xk-exk-xk+1, ak=12exk-1-xk2  and  bk+1=-12xk′. ak′=ak(bk+1-bk)  and  bk+1′=2(ak+12-ak2),  a0=0.

Furthermore, the matrix A does not change.

Proof. Let the symmetric Jacobi matrix be associated be associated with the Toda (1.9–1.11) and let J = LU, where L and U are defined by Equations (2.2, 2.3, 2.10). Consequently, the symmetric Jacobi matrix J(p) = UL is the Darboux transformation without parameter of J. By Equation (2.9), we obtain J+=J+(p), J-=J-(p) and the matrix A does not change in the Lax pair, i.e.

A=J+-J-=J+(p)-J-(p).

Moreover, similar to Equation (1.9–1.11), the symmetric Jacobi matrix J(p) = UL is associated with the Toda lattice (2.18–2.20). This completes the proof.

 3 Darboux transformation with parameter of the Jacobi matrix

The next step is the Darboux transformation with parameter of the symmetric Jacobi matrix J. We study the transformations of the polynomials of the first kind, m–functions, measure, moment sequence and Toda lattice, which are associated with the transformed Jacobi matrix.

3.1 𝔘𝔏–factorization

Theorem 3.1. Let J be the symmetric Jacobi matrix and let S0 be a some real parameter. Then J admits the following 𝔘𝔏–factorization

J=𝔘𝔏,

where 𝔏 and 𝔘 are lower and upper triangular matrices, respectively, which are defined by

𝔏=(1S0+b0a11S1+b1a21⋱⋱)  and  𝔘=(-S0a1-S1a2-S2⋱⋱),

if and only if the following system is solvable

Si(Si-1+bi-1)=-ai2,  Si-1+bi-1≠0  and  Si-1≠0,   for  all  i∈ℕ.

Proof. Let J be the Jacobi matrix. Let 𝔏 and 𝔘 are defined by Equation (3.2), where the parameter S0 ∈ ℝ \ {0, −b0}. Calculating the product 𝔘𝔏, we obtain

𝔘𝔏=(-S0a1-S1a2-S2⋱⋱)(1S0+b0a11S1+b1a21⋱⋱)=(b0a1-S1(S0+b0)a1b1a2-S2(S1+b1)a2b2⋱⋱⋱)

Comparing the product 𝔘𝔏 with the Jacobi matrix J, we obtain the system (3.3). This completes the proof.

 3.2 Transformed Jacobi matrix <italic>J</italic><sup>(<italic>d</italic>)</sup> = 𝔘𝔏

Definition 3.2. Let the symmetric Jacobi matrix J admit 𝔘𝔏—factorization of the form (3.1–3.3). Then a transformation

J=𝔘𝔏→𝔏𝔘=J(d)

is called a Darboux transformation with parameter of the Jacobi matrix J.

Theorem 3.3. Let the symmetric Jacobi matrix J admit 𝔘𝔏—factorization of the form (3.1–3.3) with parameter S0 ∈ ℝ\{0, −b0}. Then the Darboux transformation with parameter of the Jacobi matrix J is the symmetric Jacobi matrix

J(d)=(-S0a1a1b0a2a2b1⋱⋱⋱)

if and only if

S0=Si   for all  i∈ℕ.

Proof. Let J admit 𝔘𝔏—factorization of the form (3.1–3.3). Calculating the product 𝔏𝔘, we obtain

J(d)=𝔏𝔘=(-S0a1-S0(S0+b0)a1S0+b0-S1a2-S1(S1+b1)a2S1+b1-S2⋱⋱⋱).

Hence, J(d) is the symmetric Jacobi matrix if and only if

-Si-1(Si-1+bi-1)=ai2   for all  i∈ℕ.

On the other hand, by Equation (3.3), we know

-Si(Si-1+bi-1)=ai2   for all  i∈ℕ.

Consequently, we obtain Equation (3.5). This completes the proof.

Theorem 3.4. Let the symmetric Jacobi matrix J admit 𝔘𝔏—factorization of the form (3.1–3.3) and let J(d) = 𝔏𝔘 be its Darboux transformation with parameter. Then the polynomials of the first kind transform by the Geronimus formula

P0(d)(λ)≡P0(λ)  and  Pi(d)(λ)=Pi(λ)+S0+bi-1ai·Pi-1(λ),i∈ℕ,

where Pi and Pi(d) are polynomials of the first kind associated with the matrix J and J(d), respectively.

Proof. Let J admit 𝔘𝔏—factorization of the form (3.1–3.3) and let J(d) = 𝔏𝔘 be its Darboux transformation with parameter. Then

JP(λ)=λP(λ)⇒𝔘𝔏P(λ)=λP(λ)⇒𝔏𝔘𝔏P(λ)=λ𝔏P(λ)⇒⇒J(d)P(d)(λ)=λP(d)(λ),

where

P(d)(λ)=𝔏P(λ)=(1S0+b0a11S0+b1a21⋱⋱)(P0(λ)P1(λ)P2(λ)P3(λ)⋮)=(P0(λ)P1(λ)+S0+b0a1P0(λ)P2(λ)+S1+b1a2P1(λ)P3(λ)+S2+b2a3P2(λ)⋮)=(P0(d)(λ)P1(d)(λ)P2(d)(λ)P3(d)(λ)⋮).

So, the polynomials of the first kind are transformed by the Geronimus formula and Equation (3.6) holds. This completes the proof.

Proposition 3.5. Let the symmetric Jacobi matrix J admit 𝔘𝔏—factorization of the form (3.1–3.3) and let the symmetric Jacobi matrix J(d) = 𝔏𝔘 be the Darboux transformation with parameter of J. Let s={sn}n=0∞ and s(d)={sn(d)}n=0∞ be the moment sequences associated with the matrices J and J(d), respectively. Then the moment sequence s(d)={sn(d)}n=0∞ can be found by

s0(d)=1  and  sn(d)=-S0sn-1   for all  n∈ℕ.

Proof. Let the symmetric Jacobi matrix J admit 𝔘𝔏—factorization of the form (3.1–3.3) and let the symmetric Jacobi matrix J(d) = 𝔏𝔘 be its Darboux transformation with parameter. By Equation (1.5), we obtain

s0(d)=(e0,(J(d))0e0)=(e0,e0)=1

and

sn(d)=(e0,(J(d))ne0)=(e0,𝔏𝔘ne0)=(e0,𝔏(𝔏𝔘)n-1𝔘e0)=(𝔏Te0,(J)n-1(-S0)e0)=-S0(e0,(J)n-1e0)=-S0sn-1 for all  n∈ℕ.

Hence, Equation (3.7) holds. This completes the proof.

Corollary 3.6. Let the symmetric Jacobi matrix J admit 𝔘𝔏—factorization of the form (3.1–3.3) and let the symmetric Jacobi matrix J(d) = 𝔏𝔘 be the Darboux transformation with parameter of J. Let s={sn}n=0∞ and s(d)={sn(d)}n=0∞ be the moment sequences associated with the matrices J and J(d), respectively. Then

s1(d)=-S0.

Proof. By Equation (3.7) and s0 = 1, we obtain

s1(d)=-S0s0⇒s1(d)=-S0.

So, Equation (3.8) holds. This completes the proof.

Corollary 3.7. Let the symmetric Jacobi matrix J admit 𝔘𝔏—factorization of the form (3.1–3.3) and let the symmetric Jacobi matrix J(d) = 𝔏𝔘 be the Darboux transformation with parameter of J. Let 𝔖 and 𝔖(d) be the linear functionals associated with the matrices J and J(d), respectively. Then

𝔖(d)(p(λ))=-S0𝔖(p(λ)-p(0)λ)+p(0),  p(λ)∈ℂ[λ].

Proof. Let 𝔖 and 𝔖(d) be the linear functionals associated with the symmetric Jacobi matrices J = 𝔘𝔏 and J(d) = 𝔏𝔘, respectively, where 𝔏 and 𝔘 are defined by Equations (3.2, 3.3). By Equation (1.1), we obtain

𝔖(d)(λn)=sn(d)=-S0sn-1=-S0𝔖(λn-1),   for all  n∈ℕ.

Consequently, Equation (3.9) holds. This completes the proof.

Corollary 3.8. Let the symmetric Jacobi matrix J admit 𝔘𝔏—factorization of the form (3.1–3.3). and let the symmetric Jacobi matrix J(d) = 𝔏𝔘 be the Darboux transformation with parameter of J. Let dμ and dμ(d) be the measures associated with the matrices J and J(d), respectively. Then

dμ(λ)=-λS0dμ(d)(λ).

Proof. Let J = 𝔘𝔏 and J(d) = 𝔏𝔘, where the matrices 𝔏 and 𝔘 are defined by Equations (3.2, 3.3, 3.5). The measures dμ and dμ(d) are associated with the matrices J and J(d), respectively. Then

-S0∫-∞+∞λn-1dμ(λ)=-S0sn-1=sn(d)=∫-∞+∞λndμ(d)(λ).

Consequently,

∫-∞+∞λn-1dμ(λ)=-∫-∞+∞λn-1λS0dμ(d)(λ).

Hence, Equation (3.10) holds. This completes the proof.

Proposition 3.9. Let the symmetric Jacobi matrix J admit 𝔘𝔏—factorization of the form (3.1–3.3) and let the symmetric Jacobi matrix J(d) = 𝔏𝔘 be the Darboux transformation with parameter of J. Let m and m(d) be m–functions associated with the matrices J and J(d), respectively. Then

m(d)(z)=1z+S0m(z)z.

Proof. Let J = 𝔘𝔏 and J(d) = 𝔏𝔘, where the matrices 𝔏 and 𝔘 are defined by Equations (3.2, 3.3, 3.5). Then m–functions of the matrices J and J(d) are related by

m(z)=∫-∞+∞dμ(λ)λ-z=-1S0∫-∞+∞λdμ(d)(λ)λ-z=-1S0∫-∞+∞λ-zλ-zdμ(d)(λ)+1S0∫-∞+∞zdμ(d)(λ)λ-z=-s0(d)S0+zm(d)(z)S0.

On the other hand

zm(d)(z)S0=m(z)+s0(d)S0⇒m(d)(z)=s0(d)z+S0m(z)z.

By Equation (3.7), s0(d)=1 and Equation (3.11) holds. This completes the proof.

Toda latice. There is the last target of our investigation.

Theorem 3.10. Let the symmetric Jacobi matrix J admit 𝔘𝔏–factorization of the form (3.1–3.3) and J be associated with the Toda lattice (1.9–1.11). Let the symmetric Jacobi matrix J(d) = 𝔏𝔘 be the Darboux transformation without parameter of J. Then J(d) is associated with the following Toda lattice

xk″(t)=exk-1-xk-exk-xk+1, ak=12exk-1-xk2,  S0=12x0′  and  bk-1=-12xk′. a0=0,  a1′=a1(b0+S0),  ak′=ak(bk-1-bk-2),-S0′=2(a12-a02)  and  bk-1′=2(ak+12-ak2),  k∈ℕ.

Furthermore, the matrix A does not change.

Proof. Let the symmetric Jacobi matrix J be associated with the Toda lattice (1.9–1.11) and let J = 𝔘𝔏, where 𝔏 and 𝔘 are defined by Equations (3.2, 3.3, 3.5). Consequently, the symmetric Jacobi matrix J(d) = 𝔏𝔘 is the Darboux transformation with parameter of J. By Equation (3.4), we obtain J+=J+(d), J-=J-(d) and the matrix A does not change in the Lax pair, i.e.

A=J+-J-=J+(d)-J-(d).

Moreover, similar to Equations (1.9–1.11), the symmetric Jacobi matrix J(d) = 𝔏𝔘 is associated with the Toda lattice (3.12–3.14). This completes the proof.

 Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

 Author contributions

IK: Investigation, Writing – original draft. OL: Investigation, Writing – original draft.

 Funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. IK gratefully acknowledges financial support by the German Research Foundation (DFG, grant 520952250).

 Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

 Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 References Akhiezer NI. The Classical Moment Problem. Edinburgh: Oliver and Boyd (1965). Simon B. The classical moment problem as a self-adjoint finite diference operator. Adv Math. (1998) 137:82–203. 10.1006/aima.1998.1728 Berezansky YM Dudkin ME. Jacobi Matrices and the Moment Problem, Springer, Series Operator Theory: Advances and Applications. Cham: Birkhauser (2023), IX+487. 10.1007/978-3-031-46387-7 Berezanski YM. The integration of semi-infinite Toda chain by means of inverse spectral problem. Rep Math Phys. (1986) 24:21–47. 10.1016/0034-4877(86)90038-8 Gesztesy F Simon B. m-functions and inverse spectral analysis for finite and semi-infinite Jacobi matrices. J Anal Math. (1997) 73:267–97. 10.1007/BF02788147 Zhedanov A. Rational spectral transformations and orthogonal polynomials. J Comput Appl Math Vol. (1997) 85:67–86. 10.1016/S0377-0427(97)00130-1 Suetin PK. Classical Orthogonal Polynomials, 2nd rev. ed. Moscow: Nauka (1979). Toda M. Theory of Nonlinear Lattices. Springer Series in Solid-State Sciences, 20, 2nd ed. Berlin: Springer (1989). 10.1007/978-3-642-83219-2 Van Assche W. Orthogonal polynomials, toda lattices and Painlevé equations. Physica D. (2022) 434:133214. 10.1016/j.physd.2022.133214 Berezansky YM Dudkin ME editors. Applications of the spectral theory of Jacobi matrices and their generalizations to the integration of nonlinear equations. In: Jacobi Matrices and the Moment Problem Operator Theory: Advances and Applications, Vol. 294. Cham: Birkhauser (2023). p. 413–65. 10.1007/978-3-031-46387-7_9 Teschl G. Jacobi Operator and Completely Integrable Nonlinear Lattices. Providence, RI: American Mathematical Society (2001). Bueno MI Marcellán F. Darboux transformation and perturbation of linear functionals. Linear Algebra Appl. (2004) 384:215–42. 10.1016/j.laa.2004.02.004 Derevyagin M Derkach V. Darboux transformations of Jacobi matrices and Páde approximation. Linear Algebra Appl. (2011) 435:3056–84. 10.1016/j.laa.2011.05.035 Kovalyov I. Darboux transformation of generalized Jacobi matrices. Methods Funct Anal Topol. (2014) 20:301–20. Kovalyov I. Darboux transformation with parameter of generalized Jacobi matrices. J Math Sci. (2017) 222:703–22. 10.1007/s10958-017-3326-3