论文信息 - On the Problem of Decoupling Multivariate Polynomials

On the Problem of Decoupling Multivariate Polynomials

In this paper we address the application properties of the decoupling multivariate polynomials problem algorithm proposed in [2]. By numerous examples we demonstrate that this algorithm, unfortunately, fails to provide a solution in some cases. Therefore we empirically determine the application scope of this algorithm and show that it is connected with the uniqueness conditions of the CP-decomposition (Canonical Polyadic Decomposition). We also investigate the approximation properties of this algorithm and show that it is capable of construction the best low-rank polynomial approximation provided that the CP-decomposition is unique.

[1] Pierre Comon,et al. Tensors : A brief introduction , 2014, IEEE Signal Processing Magazine.

[2] Andrzej Cichocki,et al. Tensor Decompositions for Signal Processing Applications: From two-way to multiway component analysis , 2014, IEEE Signal Processing Magazine.

[3] Jan Swevers,et al. Identification of nonlinear systems using Polynomial Nonlinear State Space models , 2010, Autom..

[4] E. Bai,et al. Block Oriented Nonlinear System Identification , 2010 .

[5] Johan Schoukens,et al. Decoupling Multivariate Polynomials Using First-Order Information and Tensor Decompositions , 2014, SIAM J. Matrix Anal. Appl..

[6] J. Kruskal. Three-way arrays: rank and uniqueness of trilinear decompositions, with application to arithmetic complexity and statistics , 1977 .

[7] Johan Schoukens,et al. Data driven discrete-time parsimonious identification of a nonlinear state-space model for a weakly nonlinear system with short data record , 2018, ArXiv.

[8] J. Kruskal. Rank, decomposition, and uniqueness for 3-way and n -way arrays , 1989 .

[9] Lieven De Lathauwer,et al. A Link between the Canonical Decomposition in Multilinear Algebra and Simultaneous Matrix Diagonalization , 2006, SIAM J. Matrix Anal. Appl..