Sparse Relaxations for Coalition Structure Generation
Research Paper#Game Theory, Algorithm Design, Coalition Formation🔬 Research|Analyzed: Jan 4, 2026 00:20•
Published: Dec 25, 2025 12:56
•1 min read
•ArXivAnalysis
This paper investigates efficient algorithms for the coalition structure generation (CSG) problem, a classic problem in game theory. It compares dynamic programming (DP), MILP branch-and-bound, and sparse relaxation methods. The key finding is that sparse relaxations can find near-optimal coalition structures in polynomial time under a specific random model, outperforming DP and MILP algorithms in terms of anytime performance. This is significant because it provides a computationally efficient approach to a complex problem.
Key Takeaways
- •Compares different algorithmic paradigms for the coalition structure generation (CSG) problem.
- •Proves that sparse relaxations can achieve near-optimal solutions in polynomial time under a specific model.
- •Demonstrates a probabilistic anytime separation in favor of sparse relaxations compared to DP and MILP.
- •Highlights the computational efficiency of sparse relaxations for a complex problem.
Reference / Citation
View Original"Sparse relaxations recover coalition structures whose welfare is arbitrarily close to optimal in polynomial time with high probability."