no. 10 · Friday 6 November 2026
Cubic Optimization over Three Nonnegative Integers Is NP-hard
Abstract
We prove that deciding whether an integer cubic polynomial takes a value at most one over three nonnegative integers is NP-hard. Hardness holds even when every feasible objective value is a positive integer, so every constructed instance has a finite, attained minimum. This establishes a sharp dimension boundary for the tractability of cubic integer optimization in fixed dimension. Our reduction from positive 1-in-3 satisfiability encodes Boolean constraints as modular conditions and uses an algebraic norm to construct a cubic objective whose minimum is one exactly for satisfiable formulas. The resulting six-parameter threshold language is NP-complete and admits short exponent certificates.
- comments
- 16 pages. A version of this manuscript was posted on Optimization Online on November 1, 2026
- fields
- Optimization and Control