I know that:

$\lambda_1 , \lambda_2 , \lambda_3 \in \Re$ counting multiplicity

Tr(A) = $\lambda_1 + \lambda_2 + \lambda_3$


Det(A) = $\lambda_1 * \lambda_2 * \lambda_3$

I also know that a matrix is non-diagonalizable when the geometric and algebraic multiplicity of it's eigenvalues don't coincide. But I am unsure what to do with this information, any help would be appreciated


First note that we cannot have three different eigenvalues -- if that was the case, we would have three linearly independent eigenvectors, which would form a basis in which your matrix would diagonalize.

Second, we cannot have only one different eigenvalue: $\det(A) = \lambda^3 = 2$ would force $\lambda = \sqrt[3]{2}$, but then $Tr(a) = 3\lambda \ne 1$.

So, we need to have two different eigenvalues. Try to find out what these would be, and then try creating a matrix that is known to not be diagonalizable with these eigenvalues. Hint: do you know any non-diagonalizable matrix? Hint 2: think about Jordan blocks.


You have the following constraints on the eigenvalues.

$\lambda_1 + \lambda_2+\lambda_3 = 1\\ \lambda_1\lambda_2\lambda_3 = 2\\ \lambda_1 = \lambda_2$


$\lambda_1^2\lambda_3 = 2\\ \lambda_3 = \frac {2}{\lambda_1^2}\\ 2\lambda_1 + \frac {2}{\lambda_1^2} = 1\\ 2\lambda_1^3 - \lambda_1^2 + 2 = 0$

Now you need the cubic formula.

$\lambda_1 = \frac 16(1-\sqrt[3]{107-6\sqrt {318}}+\sqrt[3]{107+6\sqrt {318}})\approx -0.85809\\ \lambda_2 \approx -0.85809\\ \lambda_3\approx 2.71618$


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.