1
$\begingroup$

I've been working on a problem in which I routinely need to check whether a system of 4 homogeneous linear equations has a unique (zero) solution, or infinite. As of now, I'm doing it by calculating the determinant of the associated matrix.

Out of curiosity I was wondering, not being very proficient in mathematics: are there other ways to determine whether equations are independent or not? Possibly quicker than calulating the determinant, since I don't actually need any more information about the solution(s)?

Usually, the coefficients I'm working with make it impossible to just intuitively check for dependency.

$\endgroup$
2
  • $\begingroup$ If your equations are over a finite field (for coding, say), there are only finitely many solutions anyway without calculating anything. $\endgroup$ Commented May 9, 2019 at 13:54
  • $\begingroup$ Determinants are a poor way to test for dependency - the calculation can be numerically unstable. A better strategy is to calculate a row reduced form for the coefficient matrix. That's standard linear algebra. en.wikipedia.org/wiki/Gaussian_elimination $\endgroup$ Commented May 9, 2019 at 13:55

1 Answer 1

2
$\begingroup$

For larger $n$, the Gaussian elimination method requires less operations. (For manual computation, Bareiss might be more attractive).

For $n=4$, that makes little difference. Standard computation involves $24$ terms that are product of $4$ factors.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .