The Setting
Let be an ideal in a polynomial ring in variables over a field, and suppose an integer point satisfies a system of integer equations whose Jacobian minor does not vanish at . Given positive coordinate weights , a scale , and exponents , take a finite set of primes not dividing that minor at whose product is at least . The weighted residue cut asks for a bounded-degree polynomial outside vanishing on all weighted-box solutions congruent to modulo , together with a corresponding drop in dimension. Alongside these cuts stands a formal power series deformation theorem for roots with invertible Jacobian.
Polynomial Spaces and Hilbert Functions
Let be a finite variable index set, let be a field, and write for the polynomial ring over in variables , with denoting the total degree of a polynomial . For a finite variable set , a field , and a nonnegative integer , the space of polynomials of bounded degree is the -subspace of consisting of polynomials supported on monomials of total degree at most .✦ When for , write and ; for an ideal , the affine Hilbert polynomial is the univariate polynomial agreeing with the -dimension of the image of in for large integers .✦ For an ideal with affine Hilbert polynomial , the affine dimension is the degree of ✦, and the affine degree is times the coefficient of in .✦
Given a finite variable set , a real weight function , and a monomial exponent vector , the monomial weight is the real number given by (1): (1) For and a real threshold , the weighted exponent bound assigns to each the nonnegative integer floor of ✦, and the set of weighted monomials is the finite set of exponent vectors associated with and . For an ideal , a weight function , and a real threshold , the space of weighted functions is the -subspace of spanned by the quotient classes modulo of the monomials as ranges over ✦, and the weighted Hilbert function is its -dimension (2): (2)✦
Local Patches and Point Geometries
For and a field , an integer tuple is a point on an ideal if in for every .✦ For , an integer patch in variables consists of a subset of free coordinates and a family of defining equations in indexed by the non-free coordinates.✦ The patch Jacobian minor of an integer patch in variables is the determinant (3): (3)✦ An integer patch in variables is a patch on an ideal over a field if , the image of every () in belongs to , and the image of lies outside .✦
For a nonnegative integer and a real number , the integral kernel height bound is the real number defined by (4): (4) An integer patch in variables () is a -bounded integer patch, for and , if there exists an integer with such that every defining equation () satisfies and every monomial coefficient of satisfies .✦
For , a kernel geometry in variables consists of a finite point set and a real height such that for every and every .✦ An ideal () is an -kernel parent ideal, for , if is prime, , and .✦ Given , a field , and a patch piece consisting of an ideal and an integer patch in variables, a tuple is covered by if is a point on and .✦ Given , a kernel geometry in variables, nonnegative integers and , and , a patch piece in variables over is an -physical state for at if is an -kernel parent ideal, is a patch on , is an -bounded integer patch, and covers .✦
Congruence Cells and Transition Bounds
Given , a family of finite sets of nonnegative integers, and , the set of cumulative block primes at stage is the union .✦ For , a field , a finite set , an ideal , a tuple , and a finite set , the evaluation cell is the subset of points on such that for every and every .✦
For , , and , the evaluation space is the -subspace of consisting of polynomials vanishing at every point of ✦, while for the staged evaluation space is the -subspace of consisting of polynomials of total degree at most vanishing on .✦ Given , a kernel geometry in variables, a degree sequence , , , a patch piece in variables over , a family of finite sets of nonnegative integers, and , the physical kernel is the staged evaluation space (5): (5)✦
For nonnegative integers , the kernel degree budget is defined recursively on by and .✦ The patch degree budget , for , is defined recursively on , equaling at , and at equaling the maximum of with evaluated at , and the next-patch degree for .✦ Given and a sequence , the path degree budget for is defined recursively by and (6): (6)✦
Given , a kernel geometry in variables, nonnegative integers , an ideal , and a -subspace , the kernel child list is an associated finite list of patch pieces in variables over . For , , a patch piece in variables over , and a finite list of such patch pieces, the first covering child selects a patch piece in variables over . For , a kernel geometry in variables, a degree sequence , , , a patch piece in variables over with -th path degree budget for , and a -subspace , the next physical state is the patch piece in variables over given by (7): (7)✦
The Claim
Theorem 1. (i) Let be finite index sets with , let be a field of characteristic zero, let be positive weights, and let , , , and be real numbers. There exist and such that for every , every ideal satisfying for all , every family in , every with for all , every finite set of primes with for all and , and every family in satisfying , , and for all , , , and , there exists a polynomial , with image in , satisfying (8): (8)✦
(ii) Fix , positive weights , an integer , and real numbers , , , and . There exist and such that for every , every ideal with for all , every integer patch in variables with , every with for all , every finite set of primes with for all and , and every finite set whose elements satisfy for all and , for all , and for all , there exists a polynomial satisfying (9): (9)✦
(iii) Fix , positive weights , an integer , and real numbers , , , and . There exist and such that for every , all , every kernel geometry in variables whose points satisfy for , every sequence , all with , every , every -physical state for at (where is the -th path degree budget for ) satisfying and for all , every family of finite sets of nonnegative integers, and every finite set of primes with , for all , and , the ideal component of satisfies (10): (10)✦
(iv) Let be a finite index set, let be a field, let be a family of polynomials in , let satisfy for all and , and let . Then there exists a tuple of formal power series in , with constant term for each , satisfying (11): (11)✦
In the paper, the result reads: fixing and degree bounds on a patch with satisfying the Hilbert bound for (with , the defining equations of , , and standing for , , , and ), if is covered by the patch and is a set of distinct primes with for and , then for large all points in congruent to modulo lie on a bounded-degree polynomial not vanishing identically on (allowing additional congruences). The library states this cut via ideal nonmembership rather than nonvanishing on in (8) and (9), and pairs it with the affine-dimension descent (10) and the formal power series root (11).
What Is Checked
The weighted cuts (8) and (9) are stated over finite variable index sets, characteristic-zero fields and , positive weight functions, weighted monomial spaces and weighted Hilbert functions, and integer patches with their defining equations and Jacobian minors. The descent statement (10) is stated over kernel geometries, bounded-degree polynomial spaces, affine Hilbert invariants, integral kernel height bounds, physical states, path degree budgets, cumulative block primes, evaluation cells and spaces, and physical kernels, while the deformation theorem (11) is stated over multivariate polynomial systems and formal power series over an arbitrary field. For these constructions, the library proves that for every ideal over a field , there exist a rational univariate polynomial of degree at most and a threshold such that the -dimension of the image of modulo equals its value at for all .✦