How to estimate the largest eigenvalue of followed characteristics? Let $A={a_{ij}}$.

  1. Symmetric positive definite. Real. Very sparse.
  2. Diagonal elements are all positive, and off-diagonal elements are all negative.
  3. $a_{ii} $ equals to the negative of the sum of the rest elements of that row/column. $a_{ii}=-\sum^{k=n}_{k=1}a_{ik}$, $k\neq i$.
  4. $a_{ij}$ in (-1, 1].

Is this type of matrix special enough to have its own largest eigenvalue estimation?

The reason of my question is, I have 12 example matrices have the forehead mentioned characteristics, size from 30 by 30 to 3000 by 3000, but I happen to find that their largest eigenvalue only ranges in (1, 2). Therefore I am wondering if it is due to these characteristics.


You can always apply Gershgorin's lemma: given a square matrix $A$ with elements $a_{ij}$, define radiuses

$$r_i=\sum_{j=1..n,j\ne i}|a_{ij}|.$$

Then all eigenvalues of the matrix $A$ lie in the union

$$\bigcup_{i=1..n}\{z\in \Bbb C: |z-a_{ii}|\le r_i\}.$$

In your case this results in all eigenvalues lying in the set $[0,2]$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.