Efficient Shortest Path Algorithm Using An Unique And Novel Graph Matrix Representation

TL;DR

This paper introduces a novel graph matrix representation called the neighbourhood matrix and its powers, which effectively capture shortest path distances, leading to a new efficient algorithm for shortest path computation in large unweighted graphs.

Contribution

The paper extends the neighbourhood matrix concept to powers, characterizes their entries, and develops a fast shortest path algorithm validated on synthetic and real Facebook data.

Findings

The algorithm outperforms existing shortest path algorithms in running time.

The matrix powers precisely encode shortest path distances.

Empirical validation shows high efficiency on large-scale graphs.

Abstract

The neighbourhood matrix, $\mathcal{NM}(G)$, a novel representation of graphs proposed in \cite {ALPaper} is defined using the neighbourhood sets of the vertices. The matrix also exhibits a bijection between the product of two well-known graph matrices, namely the adjacency matrix and the Laplacian matrix. In this article, we extend this work and introduce the sequence of powers of $\mathcal{NM}(G)$ and denote it by $ \mathcal{NM}^{\{l\}}, 1\leq l \leq k(G) $ where $ k(G) $ is called the \textbf{iteration number}, $ k(G)=\lceil{\log_{2} diameter(G)}\rceil$. The sequence of matrices captures the distance between the vertices in a profound fashion and is found to be useful in various applications. One of the interesting results of this article is that whenever $ \eta_{ij}^{\{l\}}=-1$, for $ 1\leq l \leq k(G) $, then $d_{G}(i,j)=2^{l}$ , where $d_{G}(i,j)$ is the shortest path distance between $ i $ and $ j $. Further, we characterize the entries of the matrices $ \mathcal{NM}^{\{l\}}$, for every $l, 1\leq l \leq k(G)$. Using this concept of the sequence of powers of neighbourhood matrix and with the aid of some of its properties, we propose an algorithm to find the shortest path between any pair of vertices in a given undirected unweighted simple graph. The proposed algorithm and the claims therein are formally validated through simulations on synthetic data and the real network data from Facebook where sampling-based computations are performed for large collection of graphs containing large-sized graph. The empirical results are quite promising with our algorithm having the best running time among all the existing well known shortest path algorithms.