Work

Row-sum Constrained Totally Positive Matrix Completion

Year: 2026
Type: article
Abstract: This paper studies the completion of partially specified matrices to totally positive (TP) matrices under the additional constraint that all rows have a common sum. While a TP completion has been extensively investigated, the effect of linear normalization constraints has not been systematically addressed. Motivated by applications in spline collocation, normalized kernel operators, and stochastic models, we develop a unified framework for a row-sum constrained TP completion of $n \times n$ matrices with a single unspecified entry. Necessary interval conditions for feasibility are obtained from the positivity of all relevant $2 \times 2$ minors. The framework is then extended to rectangular $m \times n$ matrices, yielding explicit necessary interval conditions together with an algorithm for computing feasible completions. The results provide a general and constructive approach to TP completion under linear constraints, with relevance to approximation theory, numerical analysis, and related areas.
Language: English
Open Access status: hybrid
OpenAlex ID: W7167745855
FWCI:
Cites: 14
Cited by:
Related to: 10