Toric-code estimate from the degree of the evaluation set

Let X=X(Δ)X=X(\Delta) be a complete toric variety of dimension rr, let GG be a TT-invariant Cartier divisor, let PGP_G be its associated polytope, and let MM be the lattice associated to XX. Let ψG\psi_G be the support function associated to GG. For the toric evaluation code CL=CL(G,P,X)C_L=C_L(G,P,X), write nn for its length, kk for its dimension, and dd for its minimum distance. Suppose XX is nonsingular and projective, ψG\psi_G is strictly convex, and

deg(P)>deg(Gr).\operatorname{deg}(P)>\operatorname{deg}(G^r).

Toric-code degree conjecture. If qq is sufficiently large, then every fL(G)=H0(X,O(G))f\in L(G)=H^0(X,\operatorname{\mathcal O}(G)) has at most n=deg(P)n=\operatorname{deg}(P) zeros in X(Fq)X(\mathbb{F}_q). Consequently,

kdimH0(X,O(G))=PGM,k\geq \dim H^0(X,\operatorname{\mathcal O}(G))=|P_G\cap M|,

and

dnr!PGM.d\geq n-r!\cdot |P_G\cap M|.

Moreover, if n>r!PGMn>r!|P_G\cap M|, then

dimH0(X,O(G))=PGM.\dim H^0(X,\operatorname{\mathcal O}(G))=|P_G\cap M|.

The claim gives simultaneous estimates for the dimension and minimum distance of toric codes under strict convexity and a degree inequality. The supplied text explicitly records that the analogous statement is false when XX is singular, so this formulation is refuted outside its stated nonsingularity hypothesis; no refutation of the stated nonsingular version is supplied.

Sources & referencesView supporting material

Primary source

David Joyner, “Toric codes over finite fields”, arXiv:math/0208155 (2003).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.