Cofactor Expansion and Computation Methods

Computing determinants efficiently requires strategic methods beyond the definition. Cofactor expansion, row reduction, and recognition of special forms provide practical computational tools.

DefinitionMinors and Cofactors

Let $A$ be an $n \times n$ matrix.

The minor $M_{ij}$ is the determinant of the $(n-1) \times (n-1)$ submatrix obtained by deleting row $i$ and column $j$ : $M_{ij} = \det(A_{ij})$

The cofactor $C_{ij}$ is the signed minor: $C_{ij} = (-1)^{i+j}M_{ij}$

The sign pattern follows a checkerboard: $\begin{bmatrix} + & - & + & \cdots \\ - & + & - & \cdots \\ + & - & + & \cdots \\ \vdots & \vdots & \vdots & \ddots \end{bmatrix}$

TheoremLaplace Expansion

The determinant can be computed by expanding along any row or column:

Row expansion: $\det(A) = \sum_{j=1}^n a_{ij}C_{ij}$ (expand along row $i$ )

Column expansion: $\det(A) = \sum_{i=1}^n a_{ij}C_{ij}$ (expand along column $j$ )

The choice of row or column doesn't affect the result, allowing us to choose the one with most zeros for computational efficiency.

ExampleStrategic Cofactor Expansion

Compute $\det(A)$ where: $A = \begin{bmatrix} 2 & 0 & 1 & 0 \\ 3 & 0 & 0 & 1 \\ 1 & 0 & 2 & 0 \\ 0 & 0 & 4 & 1 \end{bmatrix}$

Column 2 has all zeros, making it useless. Column 4 has three zeros, so expand along column 4: $\det(A) = -0 + 1 \cdot C_{24} + 0 - 1 \cdot C_{44}$

Only two $3 \times 3$ determinants need computation, dramatically reducing work.

DefinitionComputational Methods

Row reduction method: Use elementary row operations to reduce $A$ to upper triangular form $U$ . Track sign changes from row swaps and scaling factors. Then: $\det(A) = (\pm 1) \cdot (\text{scaling factors}) \cdot \det(U) = (\pm 1) \cdot (\text{scaling factors}) \cdot \prod_{i=1}^n u_{ii}$

This method has $O(n^3)$ complexity, much faster than cofactor expansion's $O(n!)$ .

Block matrices: If $A = \begin{bmatrix} B & C \\ 0 & D \end{bmatrix}$ where $B$ and $D$ are square, then $\det(A) = \det(B)\det(D)$ .

Remark

For hand computation of small matrices (up to $4 \times 4$ ), cofactor expansion is manageable. For larger matrices or numerical computation, row reduction is essential. Special structures (diagonal, triangular, block) should always be exploited when present.