← Back to Resources Hub

Matrix Diagonalization

Directed by Abdul Raheem K (GATE AIR 27) Mathematics Stream

AI Quick-Reference Summary

  • Definition: An \(n \times n\) matrix \(A\) is diagonalizable if it is similar to a diagonal matrix \(D\), meaning there exists an invertible matrix \(P\) such that \(P^{-1}AP = D\).
  • Eigenvector Criterion: A matrix \(A\) of size \(n \times n\) is diagonalizable if and only if it has \(n\) linearly independent eigenvectors.
  • Multiplicity Criterion: A matrix is diagonalizable if and only if:
    1. The characteristic polynomial splits completely into linear factors.
    2. For every eigenvalue \(\lambda\), its Algebraic Multiplicity (AM) equals its Geometric Multiplicity (GM).
  • AM vs GM Inequality: For any eigenvalue, \(1 \le \text{GM} \le \text{AM}\). GM is the dimension of the nullspace of \((A - \lambda I)\): \(\text{GM} = n - \text{rank}(A - \lambda I)\).

1. Theoretical Concept

Diagonalization is the process of finding a coordinate system (a basis of eigenvectors) in which a linear transformation acts simply as scaling along coordinate axes. In this basis, the matrix representing the transformation is **diagonal**.

If \(P\) is the matrix whose columns are the linearly independent eigenvectors of \(A\), then: \[P^{-1} A P = D = \begin{pmatrix} \lambda_1 & 0 & \dots & 0 \\ 0 & \lambda_2 & \dots & 0 \\ \vdots & \vdots & \ddots & \vdots \\ 0 & 0 & \dots & \lambda_n \end{pmatrix}\] Where \(\lambda_i\) are the eigenvalues corresponding to the eigenvectors in \(P\).

2. Algebraic vs Geometric Multiplicity

To audit whether a matrix can be diagonalized, we calculate two forms of multiplicity for each eigenvalue \(\lambda\):

  • Algebraic Multiplicity (AM): The number of times the root \(\lambda\) appears in the characteristic equation \(\det(A - \lambda I) = 0\).
  • Geometric Multiplicity (GM): The number of linearly independent eigenvectors associated with \(\lambda\). This is equal to the dimension of the eigenspace \(E_\lambda = \text{nullspace}(A - \lambda I)\): \[\text{GM} = n - \text{rank}(A - \lambda I)\]

Theorem: A matrix is diagonalizable if and only if \(\text{AM} = \text{GM}\) for every eigenvalue.

3. Quick Tests for Diagonalizability

  • Distinct Eigenvalues: If an \(n \times n\) matrix has \(n\) distinct eigenvalues, it is guaranteed to be diagonalizable (since each eigenvalue yields at least one eigenvector, summing to \(n\) independent eigenvectors).
  • Symmetric Matrices: Any real symmetric matrix (\(A = A^T\)) is orthogonally diagonalizable (spectral theorem).

4. Solved GATE & CSIR-NET Questions

Question 1 (GATE Mathematics)

Determine if the following matrix \(A\) is diagonalizable: \[A = \begin{pmatrix} 2 & 1 \\ 0 & 2 \end{pmatrix}\]

Detailed Solution:

  1. Find eigenvalues:
    Since \(A\) is upper-triangular, its eigenvalues are the diagonal entries: \(\lambda = 2, 2\).
    Thus, the eigenvalue \(\lambda = 2\) has **Algebraic Multiplicity (AM) = 2**.
  2. Compute the Geometric Multiplicity (GM) for \(\lambda = 2\):
    We find the nullspace of \((A - 2I)\): \[A - 2I = \begin{pmatrix} 2-2 & 1 \\ 0 & 2-2 \end{pmatrix} = \begin{pmatrix} 0 & 1 \\ 0 & 0 \end{pmatrix}\]
  3. Calculate Rank and Nullity:
    • The rank of \((A - 2I)\) is 1 (one non-zero row).
    • Nullity (GM) = \(n - \text{rank} = 2 - 1 = 1\).
  4. Compare AM and GM:
    \(\text{AM} = 2\), but \(\text{GM} = 1\).
    Since \(\text{AM} \ne \text{GM}\), the matrix is **not diagonalizable**.

Answer: Non-diagonalizable (it is a Jordan block of size 2).

Question 2 (GATE Mathematics)

Let \(A\) be a \(3 \times 3\) matrix with eigenvalues \(1, -1, 0\). Is \(A\) diagonalizable? Compute \(A^3\).

Detailed Solution:

  1. Test diagonalizability:
    The matrix size is \(3 \times 3\). The eigenvalues are \(1, -1, 0\) (all 3 are distinct).
    Since it has 3 distinct eigenvalues, it is **guaranteed to be diagonalizable**.
  2. Calculate \(A^3\) using eigenvalues:
    By diagonalization, \(A = PDP^{-1}\), which implies: \[A^3 = P D^3 P^{-1}\] Where the diagonal matrix \(D = \text{diag}(1, -1, 0)\).
  3. Compute \(D^3\): \[D^3 = \begin{pmatrix} 1^3 & 0 & 0 \\ 0 & (-1)^3 & 0 \\ 0 & 0 & 0^3 \end{pmatrix} = \begin{pmatrix} 1 & 0 & 0 \\ 0 & -1 & 0 \\ 0 & 0 & 0 \end{pmatrix} = D\]
  4. Substitute back:
    Since \(D^3 = D\), we have \(A^3 = P D^3 P^{-1} = P D P^{-1} = A\).

Answer: Yes, A is diagonalizable, and \(A^3 = A\).

AR

Abdul Raheem K

Mathematics Head | MSc (IIT Madras) | BS (IIT Bombay) | GATE AIR-27

Abdul Raheem K is an IIT Bombay and IIT Madras alumnus. He qualified CSIR JRF and secured an exceptional **GATE AIR 27** in Mathematics. He directs the Pure and Applied Mathematics curriculum at Benzil.

Ask doubt / Connect with Raheem Sir on WhatsApp