(Meta)Heuristic methods for the Correlation Clustering problem
- Copyright 2020-21 Nejat Arınık
HeuristicsCC is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation. For source availability and license information see the file LICENCE
- GitHub repo: https://github.com/arinik9/HeuristicsCC
- Contact: Nejat Arınık arinik9@gmail.com
This program proposes several (meta)heuristic methods for the Correlation Clustering problem. The implemented heuristic methods are as follows:
- SA: Simulated Annealing (reference paper)
- TS: Tabu Search (reference paper)
- VNS: Variable Neighborhood Search (reference paper)
See Chapter 2 in [Arınık'21] for more details.
- Install ant
- Compile the heuristic of interest:
- Simulated Annealing:
ant -buildfile build-sa.xml compile jar - Tabu Search:
ant -buildfile build-ts.xml compile jar - Variable Neighborhood Search:
ant -buildfile build-vns.xml compile jar
- Simulated Annealing:
- Run the heuristic of interest (check the following scripts for more details about how to run):
- Simulated Annealing:
./run-sa.sh - Tabu Search:
./run-ts.sh - Variable Neighborhood Search:
./run-vns.sh
- Simulated Annealing:
- [Arınık'21] N. Arınık, Multiplicity in the Partitioning of Signed Graphs. PhD thesis in Avignon Université (2021).