Bendixson's inequality

In mathematics, Bendixson's inequality is a quantitative result in the field of matrices derived by Ivar Bendixson in 1902.[1] The inequality puts limits on the imaginary parts of Characteristic roots (Eigen values) of real matrices. A special case of this inequality leads to the result that characteristic roots of a real matrix are always real.

Mathematically, the inequality is stated as:

Let ${\displaystyle A=\left(a_{ij}\right)}$ be a real ${\displaystyle n\times n}$ matrix and ${\displaystyle \alpha =\max _{1\leq i,j\leq n}{\frac {1}{2}}\left|a_{ij}-a_{ji}\right|}$. If ${\displaystyle \lambda }$ is any characteristic root of ${\displaystyle A}$, then

${\displaystyle \left|\operatorname {Im} (\lambda )\right|\leq \alpha {\sqrt {\frac {n(n-1)}{2}}}.\,{}}$[2]

If ${\displaystyle A}$ is symmetric then ${\displaystyle \alpha =0}$ and consequently the inequality implies that ${\displaystyle \lambda }$ must be real.

References

1. An Introduction to Linear Algebra. p. 210. Retrieved 14 October 2018.
2. Iterative Solution Methods. p. 633. Retrieved 14 October 2018.