English

Acceleration for Distributed Transshipment and Parallel Maximum Flow

Data Structures and Algorithms 2025-11-20 v2

Abstract

We combine several recent advancements to solve (1+ε)(1+\varepsilon)-transshipment and (1+ε)(1+\varepsilon)-maximum flow with a parallel algorithm with O~(1/ε)\tilde{O}(1/\varepsilon) depth and O~(m/ε)\tilde{O}(m/\varepsilon) work. We achieve this by developing and deploying suitable parallel linear cost approximators in conjunction with an accelerated continuous optimization framework known as the box-simplex game by Jambulapati et al. (ICALP 2022). A linear cost approximator is a linear operator that allows us to efficiently estimate the cost of the optimal solution to a given routing problem. Obtaining accelerated ε\varepsilon dependencies for both problems requires developing a stronger multicommodity cost approximator, one where cancellations between different commodities are disallowed. For maximum flow, we observe that a recent linear cost approximator due to Agarwal et al. (SODA 2024) can be augmented with additional parallel operations and achieve ε1\varepsilon^{-1} dependency via the box-simplex game. For transshipment, we also construct a deterministic and distributed approximator. This yields a deterministic CONGEST algorithm that requires O~(ε1(D+n))\tilde{O}(\varepsilon^{-1}(D + \sqrt{n})) rounds on general networks of hop diameter DD and O~(ε1D)\tilde{O}(\varepsilon^{-1}D) rounds on minor-free networks to compute a (1+ε)(1+\varepsilon)-approximation.

Keywords

Cite

@article{arxiv.2511.06581,
  title  = {Acceleration for Distributed Transshipment and Parallel Maximum Flow},
  author = {Christoph Grunau and Rasmus Kyng and Goran Zuzic},
  journal= {arXiv preprint arXiv:2511.06581},
  year   = {2025}
}

Comments

v2: fixed typos

R2 v1 2026-07-01T07:28:42.247Z