Generalized packing–covering conjecture

For every linear code C⊆Fqn\mathcal{C}\subseteq\mathbb{F}_q^n of dimension kk over every finite field Fq\mathbb{F}_q, and every admissible order tt with 1≤t≤k1\le t\le k, the tt-th generalized Hamming weight and the tt-th generalized covering radius satisfy dt(C)≤2Rt(C)+2d_t(\mathcal{C})\le 2R_t(\mathcal{C})+2.

References

Primary source

arXiv

Additional references

Progress summary

Refreshed
Claimed solved

A computer-assisted paper claims to settle the conjecture in all cases, but its computation and proof have not yet been independently checked.

The conjecture asks whether generalized packing radii are bounded by the corresponding generalized covering radii for linear codes. A September 2026 paper claims a universal bound implying the conjecture across all finite fields, code parameters, and admissible generalized weights.

September 2026 developments

On September 28, 2026, Gianira N. Alfarano, Giuseppe Marino, Alessandro Neri, and Rocco Trombetti reported a computer-assisted proof of dt(C)≤2Rt(C)+2d_t(C)\le 2R_t(C)+2, using parity-check reformulation, puncturing, a counterexample-length bound, and exact computation. The retrieved evidence does not independently verify the implementation. Separately, Wenjun Yu and Moshe Schwartz claimed results for second-order radii, low-rate codes, and sufficiently long codes; Isaac Barouch Essayag and Aryeh Lev Zabokritskiy claimed the conjecture for redundancy at most 1414 and the bound dt(C)≤2Rt(C)+1d_t(C)\le 2R_t(C)+1 when q≥Rt(C)q\ge R_t(C).

Current status (as of September 2026): the conjecture has a claimed computer-assisted proof, while independent verification of that proof and computation is still outstanding.

Sources

Solutions 0

No solutions have been posted yet.