Concept understanding — Inverse of a Matrix by Elementary Transformations
For a non-singular square matrix A, the inverse can also be built purely from elementary row (or purely from elementary column) transformations, without ever computing a cofactor. The method rests on one rule: whenever a row transformation is applied to the prefactor of a matrix product AB=C, applying the same transformation to C keeps the equation valid while B is left untouched (the mirror rule holds for column transformations applied to a postfactor). Starting from the defining equation AA−1=I, a sequence of row transformations is applied simultaneously to the left-hand A and to the right-hand I; once the left side has been driven all the way to the identity matrix I, the right-hand side — which started as I and absorbed every transformation — has automatically become A−1. Symbolically: AA−1=Irow opsIA−1=B⇒A−1=B. If column transformations are preferred instead, the defining equation A−1A=I is used, and the transformations are applied to the columns of A and of the right-hand I, again finishing when the left side reaches I. A standard order for a 3×3 matrix is to reduce a11 to 1, then use it to clear …
Row-reducing the augmented block [A∣I] down to [I∣A−1] using elementary operations is a direct alternative to computing the adjoint and determinant separately. …
Hmm — let me re-verify this elimination arithmetic directly by instead computing A−1 via the adjugate method as a check, since the row-reduction has many steps where an arithmetic slip is easy.
Cross-check via cofactors.∣A∣=0(2⋅1−3⋅1)−1(1⋅1−3⋅3)+2(1⋅1−2⋅3)=0−1(1−9)+2(1−6)=0+8−10=−2.