Diagonal matrix scaling is NP-hard
From MaRDI portal
The author shows that the general scaling problem \(XAX e = e\), \(X = \text{diag} \{x_1,\dots,x_n\} > 0\), where \(e\) is the vector of all ones, \(X\) is a positive diagonal matrix and \(A\) is a symmetric \(n \times n\) matrix, is NP-hard. Equivalently, it is NP-hard to check for a given symmetric matrix \(A\) whether the logarithmic barrier function \({1 \over 2} x^T Ax - \Sigma \ln x_i\) has a stationary point in the positive orthant \(x > 0\).
Recommendations
Cites work
- Concerning nonnegative matrices and doubly stochastic matrices
- Diagonal Matrix Scaling and Linear Programming
- scientific article; zbMATH DE number 3639144 (Why is no real title available?)
- scientific article; zbMATH DE number 4197755 (Why is no real title available?)
- Note on Nonnegative Matrices
- On matrices with doubly stochastic pattern
- The diagonal equivalence of a nonnegative matrix to a stochastic matrix
Cited in
(7)- Finding well-conditioned similarities to block-diagonalize nonsymmetric matrices is NP-hard
- A theorem of the alternative for multihomogeneous functions and its relationship to diagonal scaling of matrices
- An information theory perspective on the balanced minimum evolution problem
- A tutorial on the balanced minimum evolution problem
- Diagonal Matrix Scaling and Linear Programming
- Tensors in computations
- Scientific contributions of Leo Khachiyan (a short overview)
This page was built for publication: Diagonal matrix scaling is NP-hard
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1908195)