betaFeedback

WorksEngineering

Tensor-Algebraic Framework for Nonlinear Control: Spectral Stability Analysis and Controller Design of Polynomial Systems

Abstract

Matrices provide an algebraic foundation for systems and control, whereas an analogous tensor-based framework for general nonhomogeneous polynomial dynamics is still lacking. This paper addresses this gap by developing a tensor-algebraic framework that connects tensor structure, spectral properties, Lyapunov stability, and controller synthesis. We first establish stability conditions based on positive-definite, copositive, and P-tensors for homogeneous and nonhomogeneous polynomial systems. We then introduce an extended tensor-eigenvalue problem (ETEP) for general multivariate polynomials, yielding a spectral-variational characterization of positivity through one interior and two boundary spectral quantities, and consequently a stability test for polynomial vector fields. For controller synthesis, we develop an unfolding-based U-eigenvalue relaxation that combines a positive-definite anchor with a near-positive-semidefinite higher-order part, leading to a convex semidefinite program. The conservatism of the relaxation is explicitly characterized, and an explicit guaranteed domain-of-attraction bound is derived. Numerical examples demonstrate positivity certification beyond direct SOS representations and substantially lower computation time than SOS-based synthesis in the considered example. These results establish a systematic tensor-spectral methodology for the analysis and control of polynomial systems with higher-order interactions.

Introduction

Matrices provide the algebraic foundation of classical systems and control theory, where structural and spectral properties are closely connected to stability, positivity, and controller design [1, 2, 3]. For instance, matrix positivity, Metzler structures, and Lyapunov inequalities lead to tractable conditions for analyzing dynamical systems [4, 5, 6]. This has established a coherent framework linking algebraic properties of matrices with the qualitative behavior of the systems they represent. Many modern systems, however, involve higher-order interactions that cannot be adequately described by pairwise matrix couplings [7]. Examples arise in social and spreading processes, biochemical and ecological networks, and multi-agent systems, where the evolution of an agent may depend simultaneously on several other components [8, 9, 10]. Such interactions naturally lead to multilinear or polynomial dynamical systems whose coefficients are more conveniently represented by tensors. In parallel, tensor algebra has developed extensions of classical matrix notions, including eigenvalues, positive definiteness, copositivity, and diagonal dominance [11, 12, 13]. These developments provide a rich algebraic foundation, but their systematic use in systems and control remains comparatively limited.

Existing tensor-based results in control and dynamical systems have mainly focused on homogeneous or specially structured settings. Tensor eigenvalues have been used to test positive definiteness of homogeneous multivariate forms, with direct motivation from Lyapunov stability analysis [14]. For homogeneous polynomial dynamical systems, orthogonally decomposable tensors enable explicit trajectory representations and sharp stability characterizations through tensor eigenpairs [15]. Tensor-based models have also been studied in application-specific settings such as hypergraph diffusion and higher-order Lotka–Volterra dynamics [8, 9]. These results demonstrate the potential of tensor methods for system analysis and control, but they largely rely on homogeneity, particular tensor decompositions, or problem-specific structures. A general tensor-spectral framework that accommodates nonhomogeneous polynomial dynamics and further supports tractable controller synthesis is still lacking.

In this paper, we consider general tensor polynomial systems of the form

x˙=Ak−1xk−1+Ak−2xk−2+⋯+A1x,\dot{x}=A_{k-1}x^{k-1}+A_{k-2}x^{k-2}+\cdots+A_1x,

where the coefficient tensors may have different orders and no special decomposition or structure is assumed. Unlike homogeneous tensor systems, the simultaneous presence of terms of different degrees prevents their stability from being characterized directly by the spectrum of a single tensor. In particular, a systematic spectral characterization of nonhomogeneous polynomial positivity, together with a computationally tractable connection to Lyapunov stability and feedback design, is still missing. Addressing this gap requires extending tensor spectral analysis beyond homogeneous forms while retaining a clear connection between algebraic tensor properties and nonlinear system behavior.

Contributions. This paper develops a unified tensor-based framework for the analysis and control of polynomial dynamical systems. First, in Sections III–VI, we derive algebraic conditions for positivity and stability of multilinear and polynomial tensor systems, and connect positive-definite, copositive, and PP-tensors with Lyapunov stability. Second, in Section VII, we establish an extended tensor–eigenvalue problem (ETEP) for general nonhomogeneous polynomials, yielding a spectral–variational positivity criterion through interior and boundary spectral quantities and a corresponding stability test for polynomial vector fields. Finally, in Sections VIII–IX, we develop an unfolding-based U-eigenvalue relaxation for controller synthesis, characterize its conservatism, and propose a convex SDP combining a low-order positive-definite anchor with a near-PSD higher-order part. An explicit guaranteed domain-of-attraction bound is further derived, providing a computationally efficient alternative to SOS-based synthesis.

Together, these results extend classical system theory from matrices to tensors and establish a comprehensive tensor-based framework for nonlinear systems with higher-order interactions. They bridge tensor algebra, spectral theory, and nonlinear control, offering new perspectives on the analysis, stabilization, and spectral characterization of polynomial dynamical systems.

Notation: Throughout the paper, R\mathbb{R} denotes the real numbers, R+\mathbb{R}_{+} and R++\mathbb{R}_{++} the nonnegative and positive reals, respectively; Rn\mathbb{R}^{n} the nn-dimensional real space; R[n,m]\mathbb{R}^{[n,m]} the space of mmth-order nn-dimensional real tensors; R+n\mathbb{R}_{+}^{n} and R++n\mathbb{R}_{++}^{n} the nonnegative and positive orthants in Rn\mathbb{R}^{n}; Sn−1={u∈Rn:∥u∥=1}\mathbb{S}^{n-1}=\{u\in\mathbb{R}^{n}:\|u\|=1\} the unit sphere; and index subsets such as I⊂{1,…,n}I\subset\{1,\ldots,n\} are used where appropriate. For M∈Rn×nM\in\mathbb{R}^{n\times n}, ρ(M)\rho(M) is its spectral radius. For M∈Rn×nM\in\mathbb{R}^{n\times n} and a∈Rna\in\mathbb{R}^{n}, MijM_{ij} and aia_i are the (i,j)(i,j)-th entry of MM and ii-th entry of aa. For a,b∈Rna,b\in\mathbb{R}^{n}, a>(<)ba>(<)b means ai>(<)bia_i>(<)b_i for all ii, and a≥(≤)ba\geq(\leq)b similarly. These comparisons extend to matrices or tensors of the same size. 1\mathbf{1} and 0\mathbf{0} denote all-ones and all-zeros with appropriate dimensions. Additional tensor-related preliminaries appear in the next section.

Preliminaries

A real tensor T∈Rn1×n2×⋯×nmT\in\mathbb{R}^{n_1\times n_2\times\cdots\times n_m} is a multidimensional array. The order of TT is mm, representing the number of its modes, and each dimension nin_i, for i=1,…,mi=1,\ldots,m, is referred to as a mode. If all modes have the same dimension nn, the tensor is called cubical and is denoted by T∈R[n,m]T\in\mathbb{R}^{[n,m]}. A cubical tensor TT is said to be supersymmetric if its entries remain invariant under any permutation of indices.

The identity tensor, denoted by I={δi1⋯im}\mathcal{I}=\{\delta_{i_1\cdots i_m}\}, is defined elementwise as

δi1⋯im={1,if i1=i2=⋯=im,0,otherwise.\delta_{i_1\cdots i_m}= \begin{cases} 1, & \text{if }i_1=i_2=\cdots=i_m,\\ 0, & \text{otherwise.} \end{cases}

Throughout this paper, unless otherwise stated, the term “tensor” refers to a cubical tensor. We now recall several types of tensor eigenvalue problems that will be used in the sequel.

Consider a tensor A∈R[n,m]A\in\mathbb{R}^{[n,m]}. The HH-eigenvalue problem [11] is defined by

Axm−1=λx[m−1],(1)Ax^{m-1}=\lambda x^{[m-1]}, \tag*{(1)}

where Axm−1Ax^{m-1} and x[m−1]x^{[m-1]} are vectors whose iith components are

(Axm−1)i=∑i2,…,im=1nAi,i2⋯imxi2⋯xim,\left(Ax^{m-1}\right)_i=\sum_{i_2,\ldots,i_m=1}^{n}A_{i,i_2\cdots i_m}x_{i_2}\cdots x_{i_m},
(x[m−1])i=xim−1.\left(x^{[m-1]}\right)_i=x_i^{m-1}.

If there exist a real scalar λ\lambda and a nonzero real vector xx satisfying (1), then λ\lambda is called an HH-eigenvalue of AA and xx the associated HH-eigenvector. If x>0x>0 (i.e., all entries are strictly positive), then λ\lambda and xx are referred to as an H++H^{++}-eigenvalue and H++H^{++}-eigenvector, respectively. In this paper we explicitly specify the type of eigenpair (H-, Z-, or E-eigenpair) whenever needed to avoid ambiguity.

In this paper, we also use the following notation: for A∈R[n,m]A\in\mathbb{R}^{[n,m]},

Axm=xTAxm−1=∑i1,…,im=1nAi1,i2⋯imxi1⋯xim.Ax^m=x^{\mathsf{T}}Ax^{m-1}=\sum_{i_1,\ldots,i_m=1}^{n}A_{i_1,i_2\cdots i_m}x_{i_1}\cdots x_{i_m}.

Another commonly used tensor eigenvalue concept is the Z-eigenvalue [11, 12]. For a tensor A∈R[n,m]A\in\mathbb{R}^{[n,m]}, the pair (λ,x)(\lambda,x) satisfies

Axm−1=λx,xTx=1.(2)Ax^{m-1}=\lambda x,\qquad x^{\mathsf{T}}x=1. \tag*{(2)}

If both λ\lambda and xx are real, they are called a Z-eigenvalue and Z-eigenvector of AA; if they are complex, they are referred to as an E-eigenvalue and E-eigenvector, respectively [11, 12].

A tensor C=(ci1⋯im)∈R[n,m]C=(c_{i_1\cdots i_m})\in\mathbb{R}^{[n,m]} is called reducible if there exists a nonempty proper subset I⊂{1,…,n}I\subset\{1,\ldots,n\} such that ci1⋯im=0c_{i_1\cdots i_m}=0, ∀i1∈I, ∀i2,…,im∉I\forall i_1\in I,\ \forall i_2,\ldots,i_m\notin I. If no such subset exists, CC is said to be irreducible. A tensor with all nonnegative entries is called a nonnegative tensor.

For a matrix R∈Rn×nR\in\mathbb{R}^{n\times n} and a tensor A∈R[n,m]A\in\mathbb{R}^{[n,m]}, their product is defined as (RA)i1i2⋯im=∑j=1nRi1jAji2⋯im(RA)_{i_1i_2\cdots i_m}=\sum_{j=1}^{n}R_{i_1j}A_{ji_2\cdots i_m}.

Following [15], an mmth-order nn-dimensional tensor A∈R[n,m]A\in\mathbb{R}^{[n,m]} is said to be almost symmetric if it is symmetric with respect to its last m−1m-1 indices. For example, a third-order tensor A∈Rn×n×nA\in\mathbb{R}^{n\times n\times n} is almost symmetric if

Ai1i2i3=Ai1i3i2,Ai2i1i3=Ai2i3i1,Ai3i1i2=Ai3i2i1.A_{i_1i_2i_3}=A_{i_1i_3i_2},\qquad A_{i_2i_1i_3}=A_{i_2i_3i_1},\qquad A_{i_3i_1i_2}=A_{i_3i_2i_1}.

Note that [15] defines almost symmetry with respect to the first m−1m-1 indices, whereas here we adopt symmetry with respect to the last m−1m-1 indices. These two definitions are equivalent, and the difference arises from the tensor–vector multiplication convention. The present notation aligns with the one used in tensor algebra [11, 16].

Lemma 1: [15] Let f:Rn→Rf:\mathbb{R}^{n}\to\mathbb{R} be a homogeneous scalar polynomial of degree mm. Then there exists a unique supersymmetric tensor AA of order mm such that f(x)=Axmf(x)=Ax^m. Similarly, for a homogeneous vector polynomial of degree m−1m-1: g:Rn→Rng:\mathbb{R}^{n}\to\mathbb{R}^{n}, there exists a unique almost symmetric tensor AA of order mm such that g(x)=Axm−1g(x)=Ax^{m-1}.

It follows directly that if AA is supersymmetric, then the derivative of f(x)=Axmf(x)=Ax^m is df(x)dx=mAxm−1\frac{df(x)}{dx}=mAx^{m-1}. If $A is almost symmetric, then the Jacobian of the vector field g(x)=Axm−1g(x)=Ax^{m-1} is ∂g(x)∂x=(m−1)Axm−2\frac{\partial g(x)}{\partial x}=(m-1)Ax^{m-2}.

Positive Tensor-based Polynomial System

In this section, we investigate the positivity properties of tensor-based dynamical systems. Recall that a dynamical system x˙=f(x)\dot{x}=f(x) is said to be positive if, for any initial condition x(0)≥0x(0)\ge0, its trajectory x(t)x(t) remains nonnegative for all t≥0t\ge0. Positivity is an essential property for systems representing population dynamics [9], opinion formation [8], or other physical processes in which states cannot become negative. We extend the classical theory of positive systems to multilinear tensor-based dynamics. Consider a general multilinear system

x˙=−Axk−1,(3)\dot{x}=-Ax^{k-1}, \tag*{(3)}

References

  1. [1]H. K. Khalil and J. W. Grizzle, Nonlinear systems. Prentice hall Upper Saddle River, NJ, 2002, vol. 3.
  2. [2]W. M. Wonham, Linear multivariable control. Springer, 1974, vol. 101.
  3. [3]F. Bullo, Lectures on Network Systems, 1.6 ed. Kindle Direct Publishing, 2022. [Online]. Available: http://motion.me.ucsb.edu/book-lns
  4. [4]K. S. Narendra and R. Shorten, “Hurwitz stability of metzler matrices,” IEEE Transactions on Automatic Control, vol. 55, no. 6, pp. 1484–1487, 2010.DOI
  5. [5]L. Gurvits, R. Shorten, and O. Mason, “On the stability of switched positive linear systems,” IEEE Transactions on Automatic Control, vol. 52, no. 6, pp. 1099–1103, 2007.
  6. [6]D. Cheng, L. Guo, and J. Huang, “On quadratic lyapunov functions,” IEEE Transactions on Automatic Control, vol. 48, no. 5, pp. 885–890, 2003.DOI
  7. [7]C. Bick, E. Gross, H. A. Harrington, and M. T. Schaub, “What are higher-order networks?” SIAM Review, vol. 65, no. 3, pp. 686–731, 2023.DOI
  8. [8]S. Cui, F. Liu, L. Liang, H. Jardón-Kojakhmetov, and M. Cao, “An sis diffusion process with direct and indirect spreading on a hypergraph,” Automatica, vol. 177, p. 112319, 2025.DOI
  9. [9]S. Cui, Q. Zhao, G. Zhang, H. Jardon-Kojakhmetov, and M. Cao, “Analysis of higher-order lotka-volterra models: Application of s-tensors and the polynomial complementarity problem,” IEEE Transactions on Automatic Control, 2025.arxiv.org/abs/2405.18333
  10. [10]S. Cui, C. Zhang, B. Jiang, H. J. Kojakhmetov, and M. Cao, “Higher-order laplacian dynamics on hypergraphs with cooperative and antagonistic interactions,” arXiv preprint arXiv:2502.08276, 2025.arxiv.org/abs/2502.08276
  11. [11]L. Qi, “Eigenvalues of a real supersymmetric tensor,” Journal of Symbolic Computation, vol. 40, no. 6, pp. 1302–1324, 2005.DOI
  12. [12]K. Chang, L. Qi, and T. Zhang, “A survey on the spectral theory of nonnegative tensors,” Numerical Linear Algebra with Applications, vol. 20, no. 6, pp. 891–912, 2013.DOI
  13. [13]Y. Wang, J. Shen, and C. Bu, “Hypergraph characterizations of copositive tensors,” Frontiers of Mathematics in China, vol. 16, no. 3, pp. 815–824, 2021.DOI
  14. [14]Q. Ni, L. Qi, and F. Wang, “An eigenvalue method for testing positive definiteness of a multivariate form,” IEEE Transactions on Automatic Control, vol. 53, no. 5, pp. 1096–1107, 2008.DOI
  15. [15]C. Chen, “Explicit solutions and stability properties of homogeneous polynomial dynamical systems,” IEEE Transactions on Automatic Control, 2022.DOI
  16. [16]L. Qi, “Symmetric nonnegative tensors and copositive tensors,” Linear Algebra and its Applications, vol. 439, no. 1, pp. 228–238, 2013.arxiv.org/abs/1211.5642
  17. [17]Z. Chen and L. Qi, “Circulant tensors with applications to spectral hypergraph theory and stochastic process,” Journal of Industrial and Management Optimization, vol. 12, no. 4, pp. 1227–1247, 2016.arxiv.org/abs/1312.2752
  18. [18]W. Ding, Z. Luo, and L. Qi, “P-tensors, p0-tensors, and their applications,” Linear Algebra and its Applications, vol. 555, pp. 336–354, 2018. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0024379518303124DOI
  19. [19]L. Chen, L. Han, T.-Y. Li, S. Li, and Z. Zhou, “Teneig - tensor eigenpairs solver.” [Online]. Available: https://users.math.msu.edu/users/chenlipi/teneig.html
  20. [20]T. S. Motzkin, “The arithmetic-geometric inequality,” Inequalities (Proc. Sympos. Wright-Patterson Air Force Base, Ohio, 1965), vol. 205, p. 54, 1967.

Paper details

vlnc:W_mucub4ty
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Engineering

ChatGPT

View statement
Looking for related papers…
Contents