Near-Optimal Approximate Shortest Paths and Transshipment in Distributed and Streaming Models

TL;DR

This paper introduces a gradient descent-based method for approximate shortest paths and transshipment problems in distributed and streaming models, achieving near-optimal efficiency without hop set constructions.

Contribution

It develops a novel gradient descent algorithm for approximate transshipment and shortest paths, improving efficiency and simplifying previous approaches in distributed and streaming settings.

Findings

Achieves $(1+\varepsilon)$-approximate SSSP in $ ilde{O}((\sqrt{n}+D)\varepsilon^{-3})$ rounds in CONGEST.

Provides $ ilde{O}(\varepsilon^{-2})$ round algorithms in congested clique and streaming models.

Circumvents hop set construction by using spanners, simplifying the process for approximate shortest path computations.

Abstract

We present a method for solving the transshipment problem - also known as uncapacitated minimum cost flow - up to a multiplicative error of $1 + \varepsilon$ in undirected graphs with non-negative edge weights using a tailored gradient descent algorithm. Using $\tilde{O}(\cdot)$ to hide polylogarithmic factors in $n$ (the number of nodes in the graph), our gradient descent algorithm takes $\tilde O(\varepsilon^{-2})$ iterations, and in each iteration it solves an instance of the transshipment problem up to a multiplicative error of $\operatorname{polylog} n$. In particular, this allows us to perform a single iteration by computing a solution on a sparse spanner of logarithmic stretch. Using a randomized rounding scheme, we can further extend the method to finding approximate solutions for the single-source shortest paths (SSSP) problem. As a consequence, we improve upon prior work by obtaining the following results: (1) Broadcast CONGEST model: $(1 + \varepsilon)$-approximate SSSP using $\tilde{O}((\sqrt{n} + D)\varepsilon^{-3})$ rounds, where $ D $ is the (hop) diameter of the network. (2) Broadcast congested clique model: $(1 + \varepsilon)$-approximate transshipment and SSSP using $\tilde{O}(\varepsilon^{-2})$ rounds. (3) Multipass streaming model: $(1 + \varepsilon)$-approximate transshipment and SSSP using $\tilde{O}(n)$ space and $\tilde{O}(\varepsilon^{-2})$ passes. The previously fastest SSSP algorithms for these models leverage sparse hop sets. We bypass the hop set construction; computing a spanner is sufficient with our method. The above bounds assume non-negative edge weights that are polynomially bounded in $n$; for general non-negative weights, running times scale with the logarithm of the maximum ratio between non-zero weights.