SmithNormalform

Smith normal form is a canonical diagonal form for integer matrices under unimodular row and column operations. For an m by n integer matrix A, there exist unimodular matrices U in GLm(Z) and V in GLn(Z) such that UAV = D, where D is diagonal with entries d1, ..., dr, followed by zeros. The positive integers satisfy d1 | d2 | ... | dr, and r is the rank of A over Z (and over Q). The nonzero di are called the invariant factors, and the pair (U, V) records the required row and column operations to reach the diagonal form. The Smith normal form is unique up to the signs of the diagonal entries, which are conventionally taken positive.

Consequences and interpretation: The Smith normal form reveals the structure of the abelian group cokernel Z^m / im(A). In particular, Z^m / im(A) ≅ Z/d1Z ⊕ ... ⊕ Z/drZ ⊕ Z^{m-r}. This provides a direct decomposition of the corresponding finitely generated abelian group and clarifies relations among the columns of A. The form also relates to solving linear Diophantine systems Ax = b, since transforming the augmented system by the same unimodular changes reduces solvability questions to simpler divisibility conditions on the diagonal entries.

Computation and applications: Smith normal form can be computed by sequences of elementary row and column operations corresponding to left and right multiplication by unimodular matrices, together with gcd computations. Algorithms exist to produce U, V and D, and implementations are available in many computer algebra systems. The Smith form is a fundamental tool in integer linear algebra, module theory over Z, and computations involving the structure of finitely generated abelian groups and chain complexes with integral coefficients.