Uniqueness and stability for the solution of a nonlinear least squares problem

DOI10.1090/MCOM/3918arXiv2104.10841OpenAlexW3155498853MaRDI QIDQ6203461FDOQ6203461

Publication date: 28 February 2024

Published in: Mathematics of Computation (Search for Journal in Brave)

Abstract: In this paper, we focus on the nonlinear least squares:

m b o x {m i n}_{m a t h b f x i n m a t h b b H^{d}} | | A m a t h b f x | - m a t h b f b |

where

A i n m a t h b b H^{m i m e s d}

,

m a t h b f b i n m a t h b b R^{m}

with

m a t h b b H i n m a t h b b R, m a t h b b C

and consider the uniqueness and stability of solutions. Such problem arises, for instance, in phase retrieval and absolute value rectification neural networks. For the case where

m a t h b f b = | A m a t h b f x_{0} |

for some

m a t h b f x_{0} i n m a t h b b H^{d}

, many results have been developed to characterize the uniqueness and stability of solutions. However, for the case where

m a t h b f b e q | A m a t h b f x_{0} |

for any

m a t h b f x_{0} i n m a t h b b H^{d}

, there is no existing result for it to the best of our knowledge. In this paper, we first focus on the uniqueness of solutions and show for any matrix

A i n m a t h b b H^{m i m e s d}

there always exists a vector

m a t h b f b i n m a t h b b R^{m}

such that the solution is not unique. But, in real case, such ``bad vectors $m a t h b f b$ are negligible, namely, if $m a t h b f b i n m a t h b b R_{+}^{m}$ does not lie in some measure zero set, then the solution is unique. We also present some conditions under which the solution is unique. For the stability of solutions, we prove that the solution is never uniformly stable. But if we restrict the vectors $m a t h b f b$ to any convex set then it is stable.

Full work available at URL: https://arxiv.org/abs/2104.10841

zbMATH Keywords

stability uniqueness nonlinear least squares Chebyshev set phase retrieval

Mathematics Subject Classification ID

Numerical mathematical programming methods (65K05) Nonconvex programming, global optimization (90C26) Information theory (general) (94A15) Sensitivity, stability, well-posedness (49K40)

Cites Work

This page was built for publication: Uniqueness and stability for the solution of a nonlinear least squares problem

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6203461)