I'm stuck with this problem, so I've got the following matrix:

$$A = \begin{bmatrix} 4& 6 & 10\\ 3& 10 & 13\\ -2&-6 &-8 \end{bmatrix}$$

Which gives me the following identity matrix of $AI$:

$$\begin{bmatrix} 4 - \lambda& 6 & 10\\ 3& 10 - \lambda & 13\\ -2&-6 & -8 - \lambda \end{bmatrix}$$

I'm looking for the Polynomial Characteristic Roots of the Determinant. I can do this on pen and paper, but I want to make this into an algorithm which can work on any given 3x3 matrix.

I can then calculate the Determinant of this matrix by doing the following:

$$Det(A) = 4 - \lambda \begin{vmatrix} 10 - \lambda&13 \\ -6 & -8 - \lambda \end{vmatrix} = \begin{bmatrix} (10 - \lambda \cdot -8 \lambda) - (-6 \cdot 13) \end{bmatrix}$$

I repeat this process for each of the columns inside the matrix (6, 10)..

Watching this video: Here the guy factorises each of the (A) + (B) + (C) to this equation:

$$ \lambda (\lambda_{2} - 6\lambda+8) = 0$$

And then finds the polynomials: 1, 2, 4.. Which I understand perfectly.

Now, putting this into code and factorising the equation would prove to be difficult. So, I'm asking whether or not there is a simple way to calculate the determinant (using the method given here) and calculate the polynomials without having to factorise the equation.. My aim is to be left with 3 roots based on the Determinant.


I think a decently efficient way to get the characteristic polynomial of a $3 \times 3$ matrix is to use the following formula: $$ P(x) = -x^3 + [tr(A)]x^2 + [[tr(A)]^2 - tr(A^2)]x + [[tr(A)]^3 + 2tr(A^3) - 3tr(A)tr(A^2)] $$ Where $tr(A)$ is the trace of $A$, and $A,A^2,A^3$ are the matrix powers of $A$.

From there, you could use the cubic formula to get the roots.

there is some computational mistake below

In this case, we'd compute $$ A = \pmatrix{4&6&10\\3&10&13\\-2&-6&-8} \implies tr(A) = 6\\ A^2 = \pmatrix{14&24&38\\16&40&56\\-10&-24&-34} \implies tr(A^2) = 20\\ A^3 = \pmatrix{52&96&148\\72&160&232\\-44&-96&-140} \implies tr(A^3) = 72 $$

  • $\begingroup$ I see where you're going with this. Just a tiny bit confusing.. What would the value of $x$ be in the case of the matrix that I gave? Also where you have $tr(A)$ is that the entire matrix, or, the $2x2% based? This sounds like a good solution, but confused over the $x$ $\endgroup$ – Phorce Jun 12 '14 at 22:30
  • $\begingroup$ I mean $x$ here the same way you used $\lambda$ in your question. $A$, $A^2$, and $A^3$ are all matrices without any variables in them. $\endgroup$ – Omnomnomnom Jun 12 '14 at 22:34
  • $\begingroup$ Hey, it kind of makes sense. However, if you get some time please can I ask if you could update your answer to give a small example of what: $A,A^2,A^3$ would look like? $\endgroup$ – Phorce Jun 12 '14 at 23:01
  • $\begingroup$ Multiply the matrix $A = \pmatrix{4&6&10\\3&10&13\\-2&-6&-8}$ with itself to get $A^2$. Multiply by $A$ again to get $A^3$. Is it not clear what I mean here? $\endgroup$ – Omnomnomnom Jun 12 '14 at 23:10
  • $\begingroup$ I see, so I will multiply matrix $A$ by itself to get $A^2$ and then take the $tr$ of that matrix. Okay! One final question before I try this on pen and paper.. Doing this equation will give me the polynomial, but, will it give me just 1 value.. In some instances, I've seen where the polynomial characteristic has been something like $\lambda -4 -5$? Thanks for the help tho :) $\endgroup$ – Phorce Jun 12 '14 at 23:13

Use the rule of Sarrus to get the characteristic polynomial and Cardano's method to get the roots...

  • $\begingroup$ We can use the Rule of Sarrus to get the polynomial? :o $\endgroup$ – Phorce Jun 12 '14 at 19:23
  • 1
    $\begingroup$ @user1326876 apply on $A-\lambda I$ to get the determinant, which is the polynominal... $\endgroup$ – draks ... Jun 12 '14 at 20:46

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.