Least-Squares Approximation by Elements from Matrix Orbits Achieved by Gradient Flows on Compact Lie Groups
Provides a unified algorithmic framework for matrix approximation problems arising in areas like signal processing and quantum information, though the approach is incremental.
The paper develops gradient-flow algorithms to find the best least-squares approximation of a given matrix by a sum of matrices from fixed orbits under unitary equivalence relations, and discusses applications across pure and applied fields.
Let $S(A)$ denote the orbit of a complex or real matrix $A$ under a certain equivalence relation such as unitary similarity, unitary equivalence, unitary congruences etc. Efficient gradient-flow algorithms are constructed to determine the best approximation of a given matrix $A_0$ by the sum of matrices in $S(A_1), ..., S(A_N)$ in the sense of finding the Euclidean least-squares distance $$\min \{\|X_1+ ... + X_N - A_0\|: X_j \in S(A_j), j = 1, >..., N\}.$$ Connections of the results to different pure and applied areas are discussed.