Polynomial reduction is the successive cancellation of polynomial terms by multiples of specified polynomials, using
a fixed monomial order. Let be a field and let
consist of nonzero polynomials
in
. If a polynomial
term
of
has monomial
divisible by the leading monomial
of
, whose leading
coefficient is
, the reduction step replaces
by
|
(1)
|
This cancels
and introduces only smaller monomials.
Repeatedly canceling the largest reducible polynomial term terminates because the monomials form a well ordered set. It yields a representation
|
(2)
|
where no monomial of the remainder is divisible by a leading
monomial of any
. The remainder can depend on
the choices of divisors. For example, with
in
and lexicographic
order
, the two reductions
|
(3)
| |||
|
(4)
|
give distinct remainders, each of which admits no further reduction by .
If
is a Gröbner basis, the remainder is unique
for the fixed monomial order and depends only on
the ideal generated by
. In that case,
belongs to the ideal iff
its remainder is zero. The coefficients
need not be unique. For an arbitrary generating set,
a zero remainder still implies membership, but a nonzero remainder does not rule
it out. For the preceding
, the polynomial
belongs to its ideal and admits no reduction by
.
Polynomial reduction is implemented in the Wolfram Language as PolynomialReduce[f,
{g1, g2, ...}, {x1, x2, ...}]. It returns the list together with the remainder
. For a single divisor in one variable, this specializes
to polynomial division.