Distributed And Parallel Low-Diameter Decompositions for Arbitrary and Restricted Graphs

Jinfeng Dou,Thorsten Götte,Henning Hillebrandt,Christian Scheideler,Julian Werthmann

2024-11-30

Abstract:We consider the distributed and parallel construction of low-diameter decompositions with strong diameter for (weighted) graphs and (weighted) graphs that can be separated through $k \in \tilde{O}(1)$ shortest paths. This class of graphs includes planar graphs, graphs of bounded treewidth, and graphs that exclude a fixed minor $K_r$. We present algorithms in the PRAM, CONGEST, and the novel HYBRID communication model that are competitive in all relevant parameters. Given $\mathcal{D} > 0$, our low-diameter decomposition algorithm divides the graph into connected clusters of strong diameter $\mathcal{D}$. For a arbitrary graph, an edge $e \in E$ of length $\ell_e$ is cut between two clusters with probability $O(\frac{\ell_e\cdot\log(n)}{\mathcal{D} })$. If the graph can be separated by $k \in \tilde{O}(1)$ paths, the probability improves to $O(\frac{\ell_e\cdot\log \log n}{\mathcal{D} })$. In either case, the decompositions can be computed in $\tilde{O}(1)$ depth and $\tilde{O}(kn)$ work in the PRAM and $\tilde{O}(1)$ time in the HYBRID model. In CONGEST, the runtimes are $\tilde{O}(HD + \sqrt{n})$ and $\tilde{O}(HD)$ respectively. All these results hold w.h.p. Broadly speaking, we present distributed and parallel implementations of sequential divide-and-conquer algorithms where we replace exact shortest paths with approximate shortest paths. In contrast to exact paths, these can be efficiently computed in the distributed and parallel setting [STOC '22]. Further, and perhaps more importantly, we show that instead of explicitly computing vertex-separators to enable efficient parallelization of these algorithms, it suffices to sample a few random paths of bounded length and the nodes close to them. Thereby, we do not require complex embeddings whose implementation is unknown in the distributed and parallel setting.

Distributed, Parallel, and Cluster Computing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: in distributed and parallel computing environments, construct a low - diameter decomposition (LDD) for any graph (including weighted graphs) and restricted graphs that can be separated by \( k \in \tilde{O}(1) \) shortest paths, and ensure that these decompositions have a strong diameter.

Specifically, the authors propose algorithms for implementing LDD in the PRAM, CONGEST, and new HYBRID communication models. These algorithms can remain competitive in all relevant parameters and work effectively for different types of graphs (such as planar graphs, bounded - treewidth graphs, and graphs excluding a fixed small clique Kr).

### Main problem decomposition

1. **Definition and goals of LDD**:
- For a given distance parameter \( D>0 \), the LDD algorithm decomposes the graph \( G=(V, E, \ell) \) into several sub - graphs (or clusters), and the strong diameter of each sub - graph does not exceed \( D \).
- The probability that each edge \( e\in E \) is cut into different clusters depends on its length \( \ell_e \), and the specific formula is:
\[
\text{For any graph:}\quad O\left(\frac{\ell_e\cdot\log(n)}{D}\right)
\]
\[
\text{For \( k \)-path - separable graphs:}\quad O\left(\frac{\ell_e\cdot\log\log n}{D}\right)
\]

2. **Applicable graph classes**:
- The paper pays special attention to a class of graphs that can be separated by \( k \in \tilde{O}(1) \) shortest paths, which include but are not limited to planar graphs, bounded - treewidth graphs, and graphs excluding a fixed small clique Kr.

3. **Algorithm performance**:
- In the PRAM model, the algorithm can be completed within a depth of \( \tilde{O}(1) \) and a workload of \( \tilde{O}(kn) \).
- In the CONGEST model, the running time is \( \tilde{O}(HD + \sqrt{n}) \) or \( \tilde{O}(HD) \).
- In the HYBRID model, the running time is \( \tilde{O}(1) \).

4. **Innovation points**:
- The paper proposes a new method to avoid explicitly computing vertex separators by sampling some random bounded - length paths and their nearby nodes, thus simplifying the implementation in distributed and parallel environments.
- The authors also show how to use approximate shortest paths instead of exact shortest paths for efficient distributed and parallel computing.

### Application scenarios

LDD plays an important role in many algorithm designs, such as:
- **Tree embedding**: It is used to construct metric tree embeddings (MTE) and low - stretch spanning trees (LSSP) to preserve the path lengths of the original graph.
- **Light spanners**: Construct light spanners so that the distances between nodes in the graph are approximately preserved.

In summary, the main contribution of this paper is to provide an efficient method for constructing distributed and parallel LDDs applicable to multiple graph classes, especially achieving near - optimal quality in \( k \)-path - separable graphs.