• Corpus ID: 14297788

A Fast Parallel Maximum Clique Algorithm for Large Sparse Graphs and Temporal Strong Components

  title={A Fast Parallel Maximum Clique Algorithm for Large Sparse Graphs and Temporal Strong Components},
  author={Ryan A. Rossi and David F. Gleich and Assefaw Hadish Gebremedhin and Md. Mostofa Ali Patwary},
We propose a fast, parallel, maximum clique algorithm for large, sparse graphs that is designed to exploit characteristics of social and information networks. We observe roughly linear runtime scaling over graphs between 1000 vertices and 100M vertices. In a test with a 1.8 billion-edge social network, the algorithm finds the largest clique in about 20 minutes. For social networks, in particular, we found that using the core number of a vertex in combination with a good heuristic clique finder… 

Figures and Tables from this paper

Coloring large complex networks

A unified framework for coloring large complex networks that consists of two main coloring variants that effectively balances the tradeoff between accuracy and efficiency is developed and is shown to be accurate with solutions close to optimal, fast and scalable for large networks, and flexible for use in a variety of applications.

Fast Algorithms for Large-Scale Network Analytics

This work provides compression, streaming, and parallel algorithms for three important graph analytics problems: centrality computation, dense subgraph discovery and community detection, and introduces new dense sub graph discovery algorithms to better model the cohesion in real-world networks.

The Shape of the Search Tree for the Maximum Clique Problem and the Implications for Parallel Branch and Bound

It is shown that load balance is sometimes a problem but that the interaction of parallel search order and the most likely location of solutions within the search space is often the dominating consideration.

Nucleus Decompositions for Identifying Hierarchy of Dense Subgraphs

The nucleus decomposition of a graph is defined, which represents the graph as a forest of nuclei, and practical algorithms for nucleus decompositions are presented and empirically evaluate their behavior in a variety of real graphs.

Finding Hierarchical and Overlapping Dense Subgraphs using Nucleus Decompositions

The nucleus decomposition of a graph is defined, which represents the graph as a forest of nuclei, which generalizes the classic notions of k-cores and k-trusses and consistently gives a global, hierarchical snapshot of dense substructures.

Cyber Graph Queries for Geographically Distributed Data Centers

New algorithms for a distributed model for graph computations motivated by limited information sharing are presented and some initial attempts to distinguish human nodes from automated nodes in social networks based only on topological properties are described.

ParK: An efficient algorithm for k-core decomposition on multicore processors

An experimental analysis of the algorithm of Batagelj and Zaversnik for k-core decomposition is presented and a new algorithm, ParK, is proposed that significantly reduces the working set size and minimizes the random accesses and is compared with state-of-the-art algorithm.

Testing a Heuristic Algorithm for Finding a Maximum Clique on DIMACS and Facebook Graphs

The proposed heuristic algorithm, called TrustCLQ, uses a general approach to solving MCP and is almost independent of the order of vertices and does not exploit a partition of the graph into independent sets.

Maximum Clique Solver Using Bitsets on GPUs

This paper presents one of the only exact maximum clique solvers that takes advantage of the parallelism of Graphical Processing Units (GPUs) and makes use of bitsets to reduce the amount of storage space needed and take advantage of bit-level parallelism in hardware to increase performance.

Finding Maximum Cliques on a Quantum Annealer

It is demonstrated that on random graphs that fit DW, no quantum speedup can be observed compared with the classical algorithms, and for instances specifically designed to fit well the DW qubit interconnection network, substantial speed-ups in computing time over classical approaches are observed.



Finding maximal cliques in massive networks

A general framework enables maximal clique enumeration to be processed recursively in small subgraphs of the input graph, thus allowing in-memory computation of maximal cliques without the costly random disk access.

Fast Algorithms for the Maximum Clique Problem on Massive Sparse Graphs

This work presents a new exact algorithm that employs novel pruning techniques and is able to quickly find maximum cliques in large sparse graphs and presents a heuristic that runs orders of magnitude faster than the exact algorithm while providing optimal or near-optimal solutions.

A scalable, parallel algorithm for maximal clique enumeration

Listing All Maximal Cliques in Large Sparse Real-World Graphs

We implement a new algorithm for listing all maximal cliques in sparse graphs due to Eppstein, Loffler, and Strash (ISAAC 2010) and analyze its performance on a large corpus of real-world graphs. Our

Listing All Maximal Cliques in Sparse Graphs in Near-optimal Time

There exists a nearly-optimal fixed-parameter tractable algorithm for enumerating all maximal cliques, parametrized by degeneracy, and this algorithm matches the Θ(d(n − d)3 d/3) worst-case output size of the problem whenever n − d = Ω(n).

The maximum clique enumeration problem: algorithms, applications, and implementations

Empirical testing reveals crucial but latent features of high-throughput biological data which distinguish real data from random data intended to reproduce salient topological features and novel decomposition strategies are tuned to the data and coupled with the best FPT MCE implementations.

Fast algorithms for maximal clique enumeration with limited memory

This work proposes an efficient partition-based algorithm for MCE that addresses the problem of processing large graphs with limited memory and reduces the high cost of CPU computation of MCE by a careful nested partition based on a cost model.

Algorithm 457: finding all cliques of an undirected graph

Two backtracking algorithms are presented, using a branchand-bound technique [4] to cut off branches that cannot lead to a clique, and generates cliques in a rather unpredictable order in an attempt to minimize the number of branches to be traversed.

Vertex neighborhoods, low conductance cuts, and good seeds for local community methods

It is theoretically demonstrate that two commonly observed properties of social networks, heavy-tailed degree distributions and large clustering coefficients, imply the existence of vertex neighborhoods (also known as egonets) that are themselves good communities.

The worst-case time complexity for generating all maximal cliques and computational experiments