find the eigenvalues

Find the Eigenvalues Fast: Step-by-Step Guide with Examples

By John White
AD

Eigenvalues are among the most revealing numbers in linear algebra. They describe how a matrix stretches, compresses, or reverses special vectors, and they appear everywhere from structural engineering and quantum mechanics to machine learning and population modeling. When you need to find the eigenvalues of a matrix, you are really uncovering the hidden scaling factors that govern its behavior.

The search begins with the eigenvalue equation Av = λv, where A is a square matrix, v is a nonzero vector, and λ is a scalar. Rearranging it as (A - λI)v = 0 shows that a nonzero solution exists only when A - λI is singular. That single condition, det(A - λI) = 0, turns the problem into a polynomial equation and gives you a dependable path to the eigenvalues.

Understanding Eigenvalues and the Characteristic Equation

The formal definition of an eigenvalue is tied to a nonzero eigenvector. If multiplying a vector by A produces the same vector scaled by λ, then λ is an eigenvalue of A and the vector is its corresponding eigenvector. The zero vector is excluded because it would make the equation true for every λ, which carries no useful information.

AD
an image of some sort of text on a black background with green and yellow lines
an image of some sort of text on a black background with green and yellow lines
an example of finding eignals and regentators in the form of numbers
an example of finding eignals and regentators in the form of numbers
a poster with the words and symbols on it
a poster with the words and symbols on it

For an n by n matrix, the characteristic equation det(A - λI) = 0 produces a polynomial of degree n. This means an n by n real matrix can have up to n eigenvalues when counted with multiplicity, and some of those eigenvalues may be complex. Repeated eigenvalues, known as algebraic multiplicity, are common in diagonal and triangular matrices, while complex eigenvalues often appear in rotation and oscillation problems.

How to Find Eigenvalues of a 2x2 Matrix

For a 2x2 matrix A = [[a, b], [c, d]], the characteristic polynomial is (a - λ)(d - λ) - bc = 0. Expanding it gives λ² - (a + d)λ + (ad - bc) = 0. The coefficient a + d is the trace, and ad - bc is the determinant, so the eigenvalues of a 2x2 matrix can be found directly from the quadratic formula.

Can You Find the Value of gf? Most People Get It Wrong!
Can You Find the Value of gf? Most People Get It Wrong!
Eigenvalues/Eigenvectors
Eigenvalues/Eigenvectors
How to Calculate Eigenvalues and Eigenvectors with NumPy - KDnuggets
How to Calculate Eigenvalues and Eigenvectors with NumPy - KDnuggets

Consider A = [[4, 1], [2, 3]]. Its trace is 7 and its determinant is 10, so the characteristic equation is λ² - 7λ + 10 = 0. Factoring gives (λ - 5)(λ - 2) = 0, so the eigenvalues are 5 and 2. You can verify each one by checking that A - λI becomes singular and by finding a nonzero vector in its null space.

Finding Eigenvalues for 3x3 and Larger Matrices

For a 3x3 matrix, det(A - λI) = 0 produces a cubic polynomial. Solving it may require factoring, the rational root theorem, or numerical methods. Special structures make the task much easier: triangular matrices have their eigenvalues on the main diagonal, diagonal matrices have them on the diagonal as well, and symmetric real matrices always have real eigenvalues.

AD

Suppose A is upper triangular with diagonal entries 2, 3, and -1. Then the eigenvalues are immediately 2, 3, and -1, because det(A - λI) is the product of the diagonal entries (2 - λ)(3 - λ)(-1 - λ). For larger matrices without a convenient structure, expanding the determinant by hand becomes impractical, so symbolic computation or numerical linear algebra is usually the better choice.

Eigenvectors and Eigenvalues explained visually
Eigenvectors and Eigenvalues explained visually
a blackboard with white writing on it in front of a wall that says find the value of y
a blackboard with white writing on it in front of a wall that says find the value of y
a book with the title and page numbers on it, which is written in english
a book with the title and page numbers on it, which is written in english
How to find function values
How to find function values
gif
gif
an image with the word find value of x and y in black ink on white paper
an image with the word find value of x and y in black ink on white paper

Practical Techniques and Numerical Methods

Finding eigenvalues can be approached symbolically or numerically. Symbolic methods return exact expressions and are ideal for small matrices, textbook exercises, and matrices with clear structure. Numerical methods return approximations and are essential for large matrices that arise in data science, engineering simulations, and scientific computing.

The best method depends on the size of the matrix, whether it is symmetric, whether you need eigenvectors as well, and how much precision you require. A reliable workflow often starts by checking for triangular, diagonal, or block structure, then moves to the characteristic polynomial or an iterative algorithm when no shortcut exists.

AD

Using the Characteristic Polynomial and Factoring

The characteristic polynomial is the foundation of exact eigenvalue computation. You compute det(A - λI), simplify the determinant, set the result equal to zero, and solve for λ. For small matrices, row operations and cofactor expansion can keep the algebra manageable, but careless arithmetic can introduce sign errors or missed factors.

Eigenvalues and Eigenvectors: How They Really Emerge
Eigenvalues and Eigenvectors: How They Really Emerge
a piece of paper with some writing on it
a piece of paper with some writing on it

If the polynomial has integer coefficients, the rational root theorem can suggest possible rational roots. For example, the matrix [[2, 0, 0], [1, 3, 0], [4, 5, -1]] is lower triangular, so its eigenvalues are 2, 3, and -1. Repeated roots require extra care because the algebraic multiplicity from the polynomial may be larger than the geometric multiplicity, which affects diagonalizability and the number of independent eigenvectors.

Numerical Algorithms: Power Iteration, QR, and Software

Power iteration finds the eigenvalue of largest magnitude by repeatedly multiplying a vector by A and normalizing it. If the matrix has a dominant eigenvalue and the starting vector has a component in the corresponding eigenvector direction, the sequence converges to that eigenvector, and the Rayleigh quotient estimates the eigenvalue. Inverse iteration applies the same idea to A inverse or a shifted matrix to target smaller or specific eigenvalues.

AD

The QR algorithm is the workhorse for dense numerical eigenvalue problems. It repeatedly factors A_k = Q_k R_k and forms A_{k+1} = R_k Q_k, gradually revealing the eigenvalues on the diagonal of an upper triangular or Schur form. Libraries such as NumPy, MATLAB, Julia, and LAPACK implement these methods with careful balancing and shifts for speed and stability. Always verify results by checking the residual Av - λv or det(A - λI), especially when eigenvalues are clustered or the matrix is ill-conditioned.

Putting these ideas into practice is the fastest way to build intuition. Start with a small matrix, compute its characteristic polynomial by hand, and confirm the eigenvalues by checking Av = λv. Then use a numerical library on a larger matrix and compare the results with any structural shortcuts you can spot. The more you connect the algebra to the geometry of scaling and rotation, the more naturally eigenvalue problems will yield their secrets.

AD