Linear Algebra: Matrices, Determinants, Eigenvalues and

A matrix is a rectangular arrangement of numbers organized into rows and columns, written inside square brackets. An m-by-n matrix has m rows and n columns. Matrices give engineers a compact way to handle many numbers at once, whether those numbers describe forces, currents, or image pixels.

Linear algebra matters because so many engineering problems boil down to systems of linear equations. Instead of solving equations one painful step at a time, we pack the coefficients into a matrix and let matrix methods do the work. The same ideas describe rotations in computer graphics, vibrations of structures, and the ranking of web pages.

Key Concepts

• Matrix and order: an m by n matrix has m rows and n columns. A square matrix has m = n.
• Identity matrix (I): ones on the diagonal, zeros elsewhere; multiplying by I leaves a matrix unchanged.
• Transpose (A^T): rows and columns swapped.
• Inverse (A^-1): the matrix such that A times A^-1 equals I. Only square matrices with non-zero determinant have an inverse.
• Determinant (det A): a single number computed from a square matrix. It is zero exactly when the matrix has no inverse.
• Eigenvalues and eigenvectors: special numbers lambda and non-zero vectors v such that A times v equals lambda times v. The matrix stretches v by the factor lambda without changing its direction.

Formulas

• Determinant of a 2 by 2 matrix: for [[a, b], [c, d]], det = ad minus bc. Dimensionless.
• Determinant of a 3 by 3 matrix: expand along any row or column using cofactors, or use the diagonal rule. Still a single scalar number.
• Matrix inverse: A^-1 = adj(A) divided by det(A), valid only when det(A) is not zero.
• Characteristic equation: det(A minus lambda I) = 0. Its roots are the eigenvalues lambda. For a 2 by 2 matrix this is a quadratic: lambda squared minus (trace) times lambda plus det = 0.
• Eigenvector equation: (A minus lambda I) v = 0, solved for v after each eigenvalue is found.

Worked Example

Find the eigenvalues of A = [[4, 1], [2, 3]].

1. Trace = 4 + 3 = 7. Determinant = (4)(3) minus (1)(2) = 12 minus 2 = 10.
2. Characteristic equation: lambda squared minus 7 lambda plus 10 = 0.
3. Factor: (lambda minus 5)(lambda minus 2) = 0, so lambda = 5 or lambda = 2.
4. For lambda = 5: (A minus 5I) v = 0 gives minus x plus y = 0, so one eigenvector is [1, 1]. Check: A times [1, 1] = [5, 5] = 5 times [1, 1]. Correct.
5. For lambda = 2: 2x plus y = 0, so one eigenvector is [1, -2].

Common Misconceptions

• Matrices multiply like ordinary numbers. Wrong: in general A times B is not equal to B times A, so order matters.
• A zero determinant means no solution. It means no unique solution: the system may have infinitely many solutions or none.
• An eigenvalue of zero makes the matrix useless. It only means the matrix is singular; the other eigenvalues and eigenvectors stay useful.

Real-World Applications

• Computer graphics: rotation, scaling, and translation of 3D objects are done with 4 by 4 matrices.
• Electrical networks: nodal and mesh analysis of large circuits are solved as matrix equations.
• Search engines: the PageRank algorithm finds the dominant eigenvector of the web link matrix.
• Structural engineering: vibration modes of buildings and bridges are eigenvectors of the stiffness matrix.

Matrices turn messy systems of equations into tidy objects with their own arithmetic. The determinant tests whether a unique solution exists, the inverse finds it, and eigenvalues with eigenvectors reveal the hidden stretching directions inside.