arXiv News

First-hand information for everyone

Site

AboutPrivacy PolicyTerms of Use

© 2026 arXiv News

arXiv News

EnglishJapanese

Switch language

EnglishJapanese
Loading account…

First-hand information for everyone

Back to front page
Data SystemsEnglishPublished

Polynomial‑time solution found for a wide class of binary affine optimization problems

November 19, 2026arXiv: 2611.22944v1

A new study shows that binary affine optimization problems whose constraint matrices have columns with at most two non‑zero entries can be solved efficiently. The authors give a polynomial‑time algorithm that reduces the task to a minimum‑weight T‑join problem on a graph built from the matrix. In contrast, when every column in the constraint matrix consistently has exactly three non‑zeros, the same problem becomes NP‑hard.

Beyond the weight‑two case, the paper develops fixed‑parameter tractable (FPT) algorithms. These algorithms run in time **O(2^rex₂(A)·max{r,m})** and **O(2^exc₂(A)·max{r,m})** respectively, where *rex₂(A)* and *exc₂(A)* measure how far the matrix is from having columns of weight at most two. The algorithms rely on a so‑called nearly‑graphic certificate that certifies this distance.

A limitation is that the FPT algorithms require such a certificate to be supplied; the work does not provide a method for finding an optimal certificate. This means that, without the certificate, the algorithms cannot be applied.

The results show a clear boundary between efficiently solvable cases and hard ones, with the transition occurring when the Hamming weight of each column increases from two to three.

Source paper

Binary Affine Optimization with Nearly-Graphic Constraint Matrices

Data SystemsMathematics
View on arXiv
Polynomial‑time solution found for a wide class of binary affine optimization problems | arXiv News