Nonsingularity, positive definiteness, and positive invertibility under fixed-point data rounding

Jiří Rohn

Applications of Mathematics (2007)

  • Volume: 52, Issue: 2, page 105-115
  • ISSN: 0862-7940

Abstract

top
For a real square matrix A and an integer d 0 , let A ( d ) denote the matrix formed from A by rounding off all its coefficients to d decimal places. The main problem handled in this paper is the following: assuming that A ( d ) has some property, under what additional condition(s) can we be sure that the original matrix A possesses the same property? Three properties are investigated: nonsingularity, positive definiteness, and positive invertibility. In all three cases it is shown that there exists a real number α ( d ) , computed solely from A ( d ) (not from A ), such that the following alternative holds: if d > α ( d ) , then nonsingularity (positive definiteness, positive invertibility) of A ( d ) implies the same property for A ; if d < α ( d ) and A ( d ) is nonsingular (positive definite, positive invertible), then there exists a matrix A ' with A ( d ) ' = A ( d ) which does not have the respective property. For nonsingularity and positive definiteness the formula for α ( d ) is the same and involves computation of the NP-hard norm · , 1 ; for positive invertibility α ( d ) is given by an easily computable formula. 0178.57901 1013.81007 0635.58034 1022.81062 0372.43005 1058.81037 0986.81031 0521.33001 0865.65009 0847.65010 0945.68077 0780.93027 0628.65027 0712.65029 0709.65036 0796.65065 0964.65049

How to cite

top

Rohn, Jiří. "Nonsingularity, positive definiteness, and positive invertibility under fixed-point data rounding." Applications of Mathematics 52.2 (2007): 105-115. <http://eudml.org/doc/33279>.

@article{Rohn2007,
abstract = {For a real square matrix $A$ and an integer $d\ge 0$, let $A_\{(d)\}$ denote the matrix formed from $A$ by rounding off all its coefficients to $d$ decimal places. The main problem handled in this paper is the following: assuming that $A_\{(d)\}$ has some property, under what additional condition(s) can we be sure that the original matrix $A$ possesses the same property? Three properties are investigated: nonsingularity, positive definiteness, and positive invertibility. In all three cases it is shown that there exists a real number $\alpha (d)$, computed solely from $A_\{(d)\}$ (not from $A$), such that the following alternative holds: $\bullet $ if $d>\alpha (d)$, then nonsingularity (positive definiteness, positive invertibility) of $A_\{(d)\}$ implies the same property for $A$; $\bullet $ if $d<\alpha (d)$ and $A_\{(d)\}$ is nonsingular (positive definite, positive invertible), then there exists a matrix $A^\{\prime \}$ with $A^\{\prime \}_\{(d)\}=A_\{(d)\}$ which does not have the respective property. For nonsingularity and positive definiteness the formula for $\alpha (d)$ is the same and involves computation of the NP-hard norm $\Vert \cdot \Vert _\{\infty ,1\}$; for positive invertibility $\alpha (d)$ is given by an easily computable formula. 0178.57901 1013.81007 0635.58034 1022.81062 0372.43005 1058.81037 0986.81031 0521.33001 0865.65009 0847.65010 0945.68077 0780.93027 0628.65027 0712.65029 0709.65036 0796.65065 0964.65049},
author = {Rohn, Jiří},
journal = {Applications of Mathematics},
keywords = {nonsingularity; positive definiteness; positive invertibility; fixed-point rounding; nonsingularity; positive definiteness; positive invertibility; fixed-point rounding},
language = {eng},
number = {2},
pages = {105-115},
publisher = {Institute of Mathematics, Academy of Sciences of the Czech Republic},
title = {Nonsingularity, positive definiteness, and positive invertibility under fixed-point data rounding},
url = {http://eudml.org/doc/33279},
volume = {52},
year = {2007},
}

TY - JOUR
AU - Rohn, Jiří
TI - Nonsingularity, positive definiteness, and positive invertibility under fixed-point data rounding
JO - Applications of Mathematics
PY - 2007
PB - Institute of Mathematics, Academy of Sciences of the Czech Republic
VL - 52
IS - 2
SP - 105
EP - 115
AB - For a real square matrix $A$ and an integer $d\ge 0$, let $A_{(d)}$ denote the matrix formed from $A$ by rounding off all its coefficients to $d$ decimal places. The main problem handled in this paper is the following: assuming that $A_{(d)}$ has some property, under what additional condition(s) can we be sure that the original matrix $A$ possesses the same property? Three properties are investigated: nonsingularity, positive definiteness, and positive invertibility. In all three cases it is shown that there exists a real number $\alpha (d)$, computed solely from $A_{(d)}$ (not from $A$), such that the following alternative holds: $\bullet $ if $d>\alpha (d)$, then nonsingularity (positive definiteness, positive invertibility) of $A_{(d)}$ implies the same property for $A$; $\bullet $ if $d<\alpha (d)$ and $A_{(d)}$ is nonsingular (positive definite, positive invertible), then there exists a matrix $A^{\prime }$ with $A^{\prime }_{(d)}=A_{(d)}$ which does not have the respective property. For nonsingularity and positive definiteness the formula for $\alpha (d)$ is the same and involves computation of the NP-hard norm $\Vert \cdot \Vert _{\infty ,1}$; for positive invertibility $\alpha (d)$ is given by an easily computable formula. 0178.57901 1013.81007 0635.58034 1022.81062 0372.43005 1058.81037 0986.81031 0521.33001 0865.65009 0847.65010 0945.68077 0780.93027 0628.65027 0712.65029 0709.65036 0796.65065 0964.65049
LA - eng
KW - nonsingularity; positive definiteness; positive invertibility; fixed-point rounding; nonsingularity; positive definiteness; positive invertibility; fixed-point rounding
UR - http://eudml.org/doc/33279
ER -

References

top
  1. Matrix Computations, The Johns Hopkins University Press, Baltimore, 1996. (1996) MR1417720
  2. Accuracy and Stability of Numerical Algorithms, SIAM, Philadelphia, 1996. (1996) Zbl0847.65010MR1368629
  3. Computational Complexity and Feasibility of Data Processing and Interval Computations, Kluwer Academic Publishers, Dordrecht, 1998. (1998) MR1491092
  4. 10.1007/BF01213466, Math. Control Signals Syst. 6 (1993), 1–9. (1993) MR1358057DOI10.1007/BF01213466
  5. Inverse-positive interval matrices, Z. Angew. Math. Mech. 67 (1987), T492–T493. (1987) Zbl0628.65027MR0907667
  6. Systems of linear interval equations, Linear Algebra Appl. 126 (1989), 39–78. (1989) Zbl0712.65029MR1040771
  7. Nonsingularity under data rounding, Linear Algebra Appl. 139 (1990), 171–174. (1990) Zbl0709.65036MR1071707
  8. 10.1137/S0895479891219216, SIAM J. Matrix Anal. Appl. 15 (1994), 175–184. (1994) Zbl0796.65065MR1257627DOI10.1137/S0895479891219216
  9. 10.1080/03081080008818644, Linear Multilinear Algebra 47 (2000), 195–204. (2000) MR1785027DOI10.1080/03081080008818644

NotesEmbed ?

top

You must be logged in to post comments.

To embed these notes on your page include the following JavaScript code on your page where you want the notes to appear.

Only the controls for the widget will be shown in your chosen language. Notes will be shown in their authored language.

Tells the widget how many notes to show per page. You can cycle through additional notes using the next and previous controls.

    
                

Note: Best practice suggests putting the JavaScript code just before the closing </body> tag.