New SSLD method enhances graph coloring heuristic with SDP preprocessing

Researchers have developed a new method called SSLD (Semidefinite Spectral Learning with DSATUR) that enhances the DSATUR heuristic for the NP-hard Graph Coloring Problem. SSLD preprocesses a graph by identifying a high-quality initial color class using Semidefinite Programming, similar to methods for computing the Lovász theta number. This preprocessing step, applied before DSATUR completes the coloring, has been shown to match or outperform DSATUR and a naive baseline across various benchmark instances, including DIMACS, random graphs, and scheduling problems. While SSLD is significantly slower than DSATUR, it demonstrates the potential of SDP-guided preprocessing for future improvements in graph coloring algorithms. AI

IMPACT This research offers a novel approach to graph coloring, potentially improving efficiency in areas like scheduling and frequency assignment where such problems arise.

RANK_REASON The cluster contains a research paper detailing a new algorithm for a computational problem. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.AI →

• arXiv
• Barabási--Albert
• DIMACS
• DSATUR
• Frequency Assignment
• Job Shop Scheduling
• Lovász theta number
• Semidefinite Programming
• Watts and Strogatz model

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →