Memetic Multilevel Hypergraph Partitioning

TL;DR

This paper introduces a novel multilevel memetic algorithm for hypergraph partitioning, significantly improving solution quality across various application domains compared to existing tools.

Contribution

It presents the first multilevel memetic approach with new recombination and mutation operations tailored for hypergraph partitioning.

Findings

Achieves superior partitioning results on benchmark instances.

Outperforms state-of-the-art tools like hMetis, PaToH, and KaHyPar.

Demonstrates effectiveness across diverse application areas.

Abstract

Hypergraph partitioning has a wide range of important applications such as VLSI design or scientific computing. With focus on solution quality, we develop the first multilevel memetic algorithm to tackle the problem. Key components of our contribution are new effective multilevel recombination and mutation operations that provide a large amount of diversity. We perform a wide range of experiments on a benchmark set containing instances from application areas such VLSI, SAT solving, social networks, and scientific computing. Compared to the state-of-the-art hypergraph partitioning tools hMetis, PaToH, and KaHyPar, our new algorithm computes the best result on almost all instances.