On nonsingular sign regular matrices. (English) Zbl 1025.15038

An \(m\times n\) real matrix \(A\) is called sign regular of order \(k\), \(1\leq k\leq \min\{m,n\}\) if for each \(l\), \(l=1,\ldots,k\), all \(l\times l\)-minors of \(A\) have the same sign \(\varepsilon_l\). The matrix \(A\) is called sign regular if \(k = \min\{m,n\}\). The list \((\varepsilon_1, \ldots,\varepsilon_k)\) of signs is called the signature of \(A\). A matrix is totally nonpositive (nonnegative) if all its minors are nonpositive (nonnegative).
The paper deals with the zero pattern of nonsingular matrices which are sign regular of order \(k\). The author shows that the amount of zero entries that can appear in a nonsingular sign regular matrix depends on its signature.
A characterization of nonsingular totally nonpositive matrices is obtained and a test of \(O(n^3)\) operations to check, if a square matrix of size \(n\) is nonsingular totally nonpositive, is provided. A key tool for the test is a Neville elimination procedure.
Nonsingular matrices with all principle minors being nonpositive (called weak \(N\)-matrices) are considered. It is known that all principal minors of a nonsingular totally nonnegative matrix are positive. However, a similar result does not hold for a nonsingular totally nonpositive matrix. It is proved that if \(A=(a_{ij})\) is a nonsingular totally nonpositive matrix with \(a_{11}<0\) and \(a_{nn}<0\) then \(A\) is a weak \(N\)-matrix.


15B57 Hermitian, skew-Hermitian, and related matrices
15B48 Positive matrices and their generalizations; cones of matrices
Full Text: DOI


[1] Ando, T., Totally positive matrices, Linear algebra appl., 90, 165-219, (1987) · Zbl 0613.15014
[2] Bapat, R.B.; Raghavan, T.E.S., Nonnegative matrices and applications, (1997), Cambridge University Press Cambridge · Zbl 0879.15015
[3] Brown, L.D.; Johnstone, I.M.; MacGibbon, K.B., Variation diminishing transformations: a direct approach to total positivity and its statistical applications, J. am. statist. assoc., 76, 824-832, (1981) · Zbl 0481.62021
[4] Cryer, C.W., LU-factorization of totally positive matrices, Linear algebra appl., 7, 83-92, (1973) · Zbl 0274.15004
[5] Fallat, S.M.; van den Driessche, P., On matrices with all minors negative, Electronic J. linear algebra, 7, 92-99, (2000) · Zbl 0962.15014
[6] Gasca, M.; Peña, J.M., Total positivity and neville elimination, Linear algebra appl., 165, 25-44, (1992) · Zbl 0749.15010
[7] Gasca, M.; Peña, J.M., A matricial description of neville elimination with applications to total positivity, Linear algebra appl., 202, 33-54, (1994) · Zbl 0804.65028
[8] Gasca, M.; Peña, J.M., A test for strict sign-regularity, Linear algebra appl., 197-198, 133-142, (1994) · Zbl 0796.65060
[9] Johnson, J.J., On partially non-positive matrices, Linear algebra appl., 8, 185-187, (1974) · Zbl 0291.15020
[10] Karlin, S., Total positivity, (1968), Stanford University Press Stanford · Zbl 0219.47030
[11] Parthasarathy, T.; Ravindran, G., N-matrices, Linear algebra appl., 139, 89-102, (1990) · Zbl 0719.15010
[12] Peña, J.M., Backward stability of a pivoting strategy for sign-regular linear systems, Bit, 37, 910-924, (1997) · Zbl 0893.65013
[13] Peña, J.M., Shape preserving representations in computer-aided geometric design, (1999), Nova Science Publishers Commack, New York · Zbl 1005.68163
[14] Saigal, R., On the class of complementary cones and lemke’s algorithm, SIAM J. appl. math., 23, 46-60, (1972) · Zbl 0237.90040
[15] Tsatsomeros, M.J.; Li, L., A recursive test for P-matrices, Bit, 40, 410-414, (2000) · Zbl 0960.65049
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.