Preprint
Article

This version is not peer-reviewed.

Research on the Yang-Baxter-Like Matrix Equation for a Diagonalizable Complex Matrix with Three Different Nonzero Eigenvalues

Submitted:

19 June 2026

Posted:

22 June 2026

You are already at the latest version

Abstract
The Yang-Baxter-like matrix equation has wide applications in various fields and holds significant theoretical research value. Suppose the coefficient matrix is an n by n diagonalizable complex matrix which has three different nonzero eigenvalues. We get some distinctive properties of solutions of the Yang-Baxter-like matrix equation.
Keywords: 
;  ;  

1. Introduction

Matrix theory is of great significance in numerous fields(see [1,2,3,4,5,6,7,8]). When matrix theory is utilized to solve practical problems, a large number of matrix equations have appeared. In this paper, we research the following equation:
A X A = X A X ,
where A is a given n × n complex matrix and X is an unknown matrix. Equation (1) is often known as Yang-Baxter-like matrix equation, or YBME for short, as it bears a strong resemblance to the classical Yang-Baxter equation
A ( u ) B ( u + v ) A ( v ) = B ( v ) A ( u + v ) B ( u ) ,
where both A and B are parametric functions of variables u and v. When A and B are independent of u and v , we can get YBME (1). The Yang-Baxter equation originates from the research on the many-body problem studied by Nobel Prize winner Yang Chen-Ning and a lattice model studied by Baxter[9,10]. It has wide applications in many fields such as quantum groups, conformal field theory and non-commutative geometry[11,12].
Equation (1) has drawn a great deal of attention on account of its extensive applications. Research on Equation (1) can be found in references [13,14,15,16,17,18], all of which presuppose that the coefficient matrix A possesses a specific structure. For example, when the coefficient matrix A is a quadratic idempotent matrix[19], a cubic idempotent matrix[20], a quartic idempotent matrix[21], or an involutory matrix[22], explicit expressions for solutions of Equation (1) are given.
Additionally, there has been extensive research on the properties of solutions of the Yang-Baxter-like matrix equation for certain special coefficient matrices. For example, when A is a matrix with nonsingular Jordan blocks, some properties of solutions of the Yang-Baxter-like matrix equation equation are given in the reference[23]
However, a general expression for solutions of Equation (1) has not been found yet. After all, solving Equation (1) is equivalent to solving n 2 quadratic equations with n 2 unknowns.
When A is a diagonalizable complex matrix which has two different eigenvalues, the explicit solutions of Equation (1) has been derived in the reference[24].
When A is a singular diagonalizable complex matrix which has three different eigenvalues, the explicit solutions of Equation (1) has been derived in the reference[?]. However, when A is a non-singular diagonalizable complex matrix which has three different nonzero eigenvalues, finding the explicit solutions of Equation (1) is very difficult.
On this basis, we research the solutions of Equation (1) when A is an n × n diagonalizable complex matrix which has three different nonzero eigenvalues and get some distinctive properties of solutions of Equation (1).
We organize the full text in the following manner. In the second section, we present the main results of this paper. In the final section, we provide a summary of the research work conducted in this paper.

2. Results

Theorem 1.
Provided that A is an n × n complex matrix. Suppose A = S J S 1 , where S is an n × n invertible complex matrix and J = λ I k μ I l β I m , λ 0 , μ 0 , β 0 , λ μ , λ β , μ β , k + l + m = n .   I k represents the k-order unit matrix, I l represents the l-order unit matrix, I m represents the m-order unit matrix. The solutions of Equation (1) are X = S Y S 1 , where Y = B C D F G H P Q W , B C k × k , C C k × l , D C k × m , F C l × k , G C l × l , H C l × m , P C m × k , Q C m × l , W C m × m . And
λ μ β P + λ ( λ μ ) P B + β ( β μ ) W P = 0 ,
λ μ β D + λ ( λ μ ) B D + β ( β μ ) D W = 0 ,
λ μ β F + λ ( λ β ) F B + μ ( μ β ) G F = 0 ,
λ μ β C + λ ( λ β ) B C + μ ( μ β ) C G = 0 .
λ μ β Q + μ ( μ λ ) Q G + β ( β λ ) W Q = 0 ,
λ μ β H + μ ( μ λ ) G H + β ( β λ ) H W = 0 ,
Proof. 
Let Y = S 1 X S . Then solving Equation (1) is equivalent to solving equation J Y J = Y J Y . We suppose Y = B C D F G H P Q W , B C k × k , C C k × l , D C k × m , F C l × k , G C l × l , H C l × m , P C m × k , Q C m × l , W C m × m . Then
J Y J = λ I k μ I l β I m B C D F G H P Q W λ I k μ I l β I m = λ 2 B λ μ C λ β D λ μ F μ 2 G μ β H λ β P μ β Q β 2 W ,
Y J Y = B C D F G H P Q W λ I k μ I l β I m B C D F G H P Q W = λ B 2 + μ C F + β D P λ B C + μ C G + β D Q λ B D + μ C H + β D W λ F B + μ G F + β H P λ F C + μ G 2 + β H Q λ F D + μ G H + β H W λ P B + μ Q F + β W P λ P C + μ Q G + β W Q λ P D + μ Q H + β W 2 .
As J Y J = Y J Y , we can get that
λ 2 B = λ B 2 + μ C F + β D P ,
λ μ C = λ B C + μ C G + β D Q ,
λ β D = λ B D + μ C H + β D W ,
λ μ F = λ F B + μ G F + β H P ,
μ 2 G = λ F C + μ G 2 + β H Q ,
μ β H = λ F D + μ G H + β H W ,
λ β P = λ P B + μ Q F + β W P ,
μ β Q = λ P C + μ Q G + β W Q ,
β 2 W = λ P D + μ Q H + β W 2 .
(a) According to Equation (16), we know
β W 2 = β 2 W λ P D μ Q H .
Further
β W 2 P = β 2 W P λ P D P μ Q H P .
According to Equation (14), we know
β W P = λ β P λ P B μ Q F .
Further
β W 2 P = λ β W P λ W P B μ W Q F .
So
β 2 W P λ P D P μ Q H P = λ β W P λ W P B μ W Q F .
In other words,
β ( β λ ) W P λ P D P μ Q H P + λ W P B + μ W Q F = 0 .
Further
β 2 ( β λ ) W P λ β P D P μ β Q H P + λ β W P B + μ β W Q F = 0 .
According to Equation (8), we know
β D P = λ 2 B λ B 2 μ C F .
Further
λ β P D P = λ 3 P B + λ 2 P B 2 + λ μ P C F .
According to Equation (11), we know
β H P = λ μ F λ F B μ G F .
Further
μ β Q H P = λ μ 2 Q F + λ μ Q F B + μ 2 Q G F .
According to equation(14), we know
β W P = λ β P λ P B μ Q F .
Further
λ β W P B = λ 2 β P B λ 2 P B 2 λ μ Q F B .
According to Equation (15), we know
β W Q = μ β Q λ P C μ Q G .
Further
μ β W Q F = μ 2 β Q F λ μ P C F μ 2 Q G F .
Substituting Equations (18), (19), (20) and (21) into Equation (17), we are able to get
β 2 ( β λ ) W P + μ 2 ( β λ ) Q F + λ 2 ( β λ ) P B = 0 .
As β λ , we can get that
β 2 W P + μ 2 Q F + λ 2 P B = 0 .
According to Equation(14), we know
μ Q F = λ β P λ P B β W P .
So
β 2 W P + μ ( λ β P λ P B β W P ) + λ 2 P B = 0 .
That is to say,
λ μ β P + λ ( λ μ ) P B + β ( β μ ) W P = 0 .
(b) According to Equation (8), we know
λ B 2 = λ 2 B μ C F β D P .
Further
λ B 2 D = λ 2 B D μ C F D β D P D .
According to Equation (10), we know
λ B D = λ β D μ C H β D W .
Further
λ B 2 D = λ β B D μ B C H β B D W .
So
λ 2 B D μ C F D β D P D = λ β B D μ B C H β B D W .
In other words,
λ ( λ β ) B D μ C F D β D P D + μ B C H + β B D W = 0 .
Further
λ 2 ( λ β ) B D λ μ C F D λ β D P D + λ μ B C H + λ β B D W = 0 .
According to Equation (13), we know
λ F D = μ β H μ G H β H W .
Further
λ μ C F D = μ 2 β C H + μ 2 C G H + μ β C H W .
According to Equation (16), we know
λ P D = β 2 W μ Q H β W 2 .
Further
λ β D P D = β 3 D W + μ β D Q H + β 2 D W 2 .
According to Equation (9), we know
λ B C = λ μ C μ C G β D Q .
Further
λ μ B C H = λ μ 2 C H μ 2 C G H μ β D Q H .
According to Equation (10), we know
λ B D = λ β D μ C H β D W .
Further
λ β B D W = λ β 2 D W μ β C H W β 2 D W 2 .
Substituting Equations (23), (24), (25) and (26) into Equation (22), we are able to get
λ 2 ( λ β ) B D + μ 2 ( λ β ) C H + β 2 ( λ β ) D W = 0 .
As λ β , we can get that
λ 2 B D + μ 2 C H + β 2 D W = 0 .
According to equation(10), we know
μ C H = λ β D λ B D β D W .
So
λ 2 B D + μ ( λ β D λ B D β D W ) + β 2 D W = 0 .
That is to say,
λ μ β D + λ ( λ μ ) B D + β ( β μ ) D W = 0 .
(c) According to Equation (12), we know
μ G 2 = μ 2 G λ F C β H Q .
Further
μ G 2 F = μ 2 G F λ F C F β H Q F .
According to Equation (11), we know
μ G F = λ μ F λ F B β H P .
Further
μ G 2 F = λ μ G F λ G F B β G H P .
So
μ 2 G F λ F C F β H Q F = λ μ G F λ G F B β G H P .
In other words,
μ ( μ λ ) G F λ F C F β H Q F + λ G F B + β G H P = 0 .
Further
μ 2 ( μ λ ) G F λ μ F C F μ β H Q F + λ μ G F B + μ β G H P = 0 .
According to Equation (8), we know
μ C F = λ 2 B λ B 2 β D P .
Further
λ μ F C F = λ 3 F B + λ 2 F B 2 + λ β F D P .
According to Equation (14), we know
μ Q F = λ β P λ P B β W P .
Further
μ β H Q F = λ β 2 H P + λ β H P B + β 2 H W P .
According to Equation (13), we know
μ G H = μ β H λ F D β H W .
Further
μ β G H P = μ β 2 H P λ β F D P β 2 H W P .
According to Equation (11), we know
μ G F = λ μ F λ F B β H P .
Further
λ μ G F B = λ 2 μ F B λ 2 F B 2 λ β H P B .
Substituting Equations (28), (29), (30) and (31) into Equation (27), we are able to get
λ 2 ( λ μ ) F B + μ 2 ( λ μ ) G F + β 2 ( λ μ ) H P = 0 .
As λ μ , we can get that
λ 2 F B + μ 2 G F + β 2 H P = 0 .
According to equation(11), we know
β H P = λ μ F λ F B μ G F .
So
λ 2 F B + μ 2 G F + β ( λ μ F λ F B μ G F ) = 0 .
That is to say,
λ μ β F + λ ( λ β ) F B + μ ( μ β ) G F = 0 .
(d) According to Equation (8), we know
λ B 2 = λ 2 B μ C F β D P .
Further
λ B 2 C = λ 2 B C μ C F C β D P C .
According to Equation (9), we know
λ B C = λ μ C μ C G β D Q .
Further
λ B 2 C = λ μ B C μ B C G β B D Q .
So
λ 2 B C μ C F C β D P C = λ μ B C μ B C G β B D Q .
In other words,
λ ( λ μ ) B C μ C F C β D P C + μ B C G + β B D Q = 0 .
Further
λ 2 ( λ μ ) B C λ μ C F C λ β D P C + λ μ B C G + λ β B D Q = 0 .
According to Equation (12), we know
λ F C = μ 2 G μ G 2 β H Q .
Further
λ μ C F C = μ 3 C G + μ 2 C G 2 + μ β C H Q .
According to Equation (15), we know
λ P C = μ β Q μ Q G β W Q .
Further
λ β D P C = μ β 2 D Q + μ β D Q G + β 2 D W Q .
According to Equation (9), we know
λ B C = λ μ C μ C G β D Q .
Further
λ μ B C G = λ μ 2 C G μ 2 C G 2 μ β D Q G .
According to Equation (10), we know
λ B D = λ β D μ C H β D W .
Further
λ β B D Q = λ β 2 D Q μ β C H Q β 2 D W Q .
Substituting Equations (33), (34), (35) and (36) into Equation (32), we are able to get
λ 2 ( λ μ ) B C + μ 2 ( λ μ ) C G + β 2 ( λ μ ) D Q = 0 .
As λ μ , we can get that
λ 2 B C + μ 2 C G + β 2 D Q = 0 .
According to equation(9), we know
β D Q = λ μ C λ B C μ C G .
So
λ 2 B C + μ 2 C G + β ( λ μ C λ B C μ C G ) = 0 .
That is to say,
λ μ β C + λ ( λ β ) B C + μ ( μ β ) C G = 0 .
(e) According to Equation (16), we know
β W 2 = β 2 W λ P D μ Q H .
Further
β W 2 Q = β 2 W Q λ P D Q μ Q H Q .
According to Equation (15), we know
β W Q = μ β Q λ P C μ Q G .
Further
β W 2 Q = μ β W Q λ W P C μ W Q G .
So
β 2 W Q λ P D Q μ Q H Q = μ β W Q λ W P C μ W Q G .
In other words,
β ( β μ ) W Q λ P D Q μ Q H Q + λ W P C + μ W Q G = 0 .
Further
β 2 ( β μ ) W Q λ β P D Q μ β Q H Q + λ β W P C + μ β W Q G = 0 .
According to Equation (9), we know
β D Q = λ μ C λ B C μ C G .
Further
λ β P D Q = λ 2 μ P C + λ 2 P B C + λ μ P C G .
According to Equation (12), we know
β H Q = μ 2 G λ F C μ G 2 .
Further
μ β Q H Q = μ 3 Q G + λ μ Q F C + μ 2 Q G 2 .
According to Equation (14), we know
β W P = λ β P λ P B μ Q F .
Further
λ β W P C = λ 2 β P C λ 2 P B C λ μ Q F C .
According to Equation (15), we know
β W Q = μ β Q λ P C μ Q G .
Further
μ β W Q G = μ 2 β Q G λ μ P C G μ 2 Q G 2 .
Substituting Equations (38), (39), (40) and (41) into Equation (37), we are able to get
λ 2 ( β μ ) P C + μ 2 ( β μ ) Q G + β 2 ( β μ ) W Q = 0 .
As β μ , we can get that
λ 2 P C + μ 2 Q G + β 2 W Q = 0 .
According to equation(15), we know
λ P C = μ β Q μ Q G β W Q .
So
λ ( μ β Q μ Q G β W Q ) + μ 2 Q G + β 2 W Q = 0 .
That is to say,
λ μ β Q + μ ( μ λ ) Q G + β ( β λ ) W Q = 0 .
(f) According to Equation (12), we know
μ G 2 = μ 2 G λ F C β H Q .
Further
μ G 2 H = μ 2 G H λ F C H β H Q H .
According to Equation (13), we know
μ G H = μ β H λ F D β H W .
Further
μ G 2 H = μ β G H λ G F D β G H W .
So
μ 2 G H λ F C H β H Q H = μ β G H λ G F D β G H W .
In other words,
μ ( μ β ) G H λ F C H β H Q H + λ G F D + β G H W = 0 .
Further
μ 2 ( μ β ) G H λ μ F C H μ β H Q H + λ μ G F D + μ β G H W = 0 .
According to Equation (10), we know
μ C H = λ β D λ B D β D W .
Further
λ μ F C H = λ 2 β F D + λ 2 F B D + λ β F D W .
According to Equation (16), we know
μ Q H = β 2 W λ P D β W 2 .
Further
μ β H Q H = β 3 H W + λ β H P D + β 2 H W 2 .
According to Equation (11), we know
μ G F = λ μ F λ F B β H P .
Further
λ μ G F D = λ 2 μ F D λ 2 F B D λ β H P D .
According to Equation (13), we know
μ G H = μ β H λ F D β H W .
Further
μ β G H W = μ β 2 H W λ β F D W β 2 H W 2 .
Substituting Equations (43), (44), (45) and (46) into Equation (42), we are able to get
λ 2 ( μ β ) F D + μ 2 ( μ β ) G H + β 2 ( μ β ) H W = 0 .
As μ β , we can get that
λ 2 F D + μ 2 G H + β 2 H W = 0 .
According to equation(13), we know
λ F D = μ β H μ G H β H W .
So
λ ( μ β H μ G H β H W ) + μ 2 G H + β 2 H W = 0 .
That is to say,
λ μ β H + μ ( μ λ ) G H + β ( β λ ) H W = 0 .

3. Conclusions

When A is an n × n complex matrix which is diagonalizable and has three different nonzero eigenvalues, we get some distinctive properties of solutions of Equation (1). However, solving Equation (1) remains a significant challenge, necessitating further research.

References

  1. H. Y. Li, S. Y. Chen, S. F. Xu, Z. M. Song, J. X. Chen, D. Zhao, EMI signal feature enhancement based on extreme energy difference and deep autoencoder, IET Signal Processing. 12 (2018), 852-856. [CrossRef]
  2. H. Y. Li, L. Li, D. Zhao, An improved EMD method with modified envelope algorithm based on C2 piecewise rational cubic spline interpolation for EMI signal decomposition, Appl. Math. Comput. 335 (2018), 112-123. [CrossRef]
  3. L. Li, H. Y. Li, D. Zhao, A Schwarz-Pick lemma for the modulus of holomorphic mappings from  B n p  to  B m p , Complex Var. Elliptic Equ. 62 (2017), 1746–1757.
  4. H. Y. Li, X. Y. Qin, D. Zhao, J. X. Chen, P. D. Wang, An improved empirical mode decomposition method based on the cubic trigonometric B-spline interpolation algorithm, Appl. Math. Comput. 332 (2018), 406–419. [CrossRef]
  5. H. Y. Li, C. J. Wang, D. Zhao, Preconditioning for PDE-constrained optimization with total variation regularization, Appl. Math. Comput. 386 (2020), 125470. [CrossRef]
  6. H. Y. Li, D. Zhao, F. Dai, D. L Su, On the spectral radius of a nonnegative centrosymmetric matrix, Appl. Math. Comput. 218 (2012), 4962–4966. [CrossRef]
  7. D. Zhao, H. Y. Li, On the computation of inverses and determinants of a kind of special matrices, Appl. Math. Comput. 250 (2015), 721–726. [CrossRef]
  8. D. Zhao, H. Y. Li, Z. S. Gao, Trace inequalities for positive semidefinite matrices with centrosymmetric structure, J. Inequal. Appl. (2012), 62. [CrossRef]
  9. R. J. Baxter,Partition function of the eight-vertex lattice model, Ann. Phys. 70(1972), 193-228.
  10. C. N. Yang, Some exact results for the many-body problem in one dimension with repulsive delta-function interaction, Phys. Rev. Lett. 19 (1967), 1312-1315. [CrossRef]
  11. T. Gateva-Ivanova, Quadratic algebras, Yang–Baxter equation, and Artin–Schelter regularity, Adv. Math. 230 (2012), 2152-2175. [CrossRef]
  12. J. Hietarinta, All solutions to the constant quantum Yang-Baxter equation in two dimensions, Phys. Lett. A. 165 (1992), 245-251. [CrossRef]
  13. D. M. Chen, Z.B. Chen, X.R. Yong, Explicit solutions of the Yang-Baxter-like matrix equation for a diagonalizable matrix with spectrum contained in {1,α,0}, Appl. Math. Comput. 348(2019), 523–530. [CrossRef]
  14. J. Ding, N. H. Rhee, Computing solutions of the Yang-Baxter-like matrix equation for diagonalisable matrices, East Asian Journal on Applied Mathematics. 5 (2015), 75–84. [CrossRef]
  15. Q. X. Dong, J. Ding, All projection-based commuting solutions of the Yang-Baxter-like matrix equation, Filomat. 35 (2021), 3203–3217. [CrossRef]
  16. D.Zhou, G.Chen, J.Ding, Solving the Yang-Baxter-like matrix equation for rank-two matrices, J. Comput. Appl. Math. 313 (2017), 142–151.
  17. A. Kumar, J. R. Cardoso, Iterative methods for finding commuting solutions of the Yang-Baxter-like matrix equation, Appl. Math. Comput. 333 (2018), 246–253. [CrossRef]
  18. L. Lu, Constructing solutions of the Yang-Baxter-like matrix equation for singular matrices, J. Comput. Appl. Math. 436 (2024), 115403. [CrossRef]
  19. S. Mansour, J.Ding, Q. Huang, Explicit solutions of the Yang-Baxter-like matrix equation for an idempotent matrix, Appl. Math. Lett. 63 (2017), 71-76.
  20. S. Mansour, J.Ding, Q.Huang, L.P.Zhu, All solutions of the Yang-Baxter-like matrix equation when A3=A, J. Appl. Anal. Comput. 9 (2019), 1022-1031.
  21. D. M. Zhou, H. Q. Vu, Some non-commuting solutions of the Yang-Baxter-like matrix equation, Open Math. 18 (2020), 948–969. [CrossRef]
  22. M. S. I. Adam, J. Ding, Q. L. Huang, L. P. Zhu, Solving a class of quadratic matrix equations, Applied Mathematics Letters. 82 (2018), 58–63. [CrossRef]
  23. S. J. Chen, X. L. Li, L. P. Zhu, Q. L. Huang, Solutions of the Yang-Baxter-like matrix equation for the matrix with nonsingular Jordan blocks, J. Appl. Anal. Comput.13 (2023), 986–999. [CrossRef]
  24. D.Shen, M. Wei, All solutions of the Yang-Baxter-like matrix equation for diagonalizable coefficient matrix with two different eigenvalues, Appl. Math. Lett. 101 (2020), 106048.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.
Prerpints.org logo

Preprints.org is a free preprint server supported by MDPI in Basel, Switzerland.

Subscribe

© 2026 MDPI (Basel, Switzerland) unless otherwise stated

Accessibility

Disclaimer

Terms of Use

Privacy Policy

Privacy Settings