HeuristicsCC

(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

---

Description

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.

Use

1. Install ant
2. Compile the heuristic of interest:
   1. Simulated Annealing: ant -buildfile build-sa.xml compile jar
   2. Tabu Search: ant -buildfile build-ts.xml compile jar
   3. Variable Neighborhood Search: ant -buildfile build-vns.xml compile jar
3. Run the heuristic of interest (check the following scripts for more details about how to run):
   1. Simulated Annealing: ./run-sa.sh
   2. Tabu Search: ./run-ts.sh
   3. Variable Neighborhood Search: ./run-vns.sh

References

• [Arınık'21] N. Arınık, Multiplicity in the Partitioning of Signed Graphs. PhD thesis in Avignon Université (2021).