Learning Combinatorial Optimization on Graphs: A Survey with Applications to Networking

TL;DR

This survey reviews machine learning methods for solving combinatorial optimization problems on graphs, emphasizing their applications in networking and addressing the limitations of traditional algorithmic approaches.

Contribution

It provides an organized comparison of learning-based techniques for graph optimization problems, focusing on practical applications in telecommunications networks.

Findings

Machine learning offers flexible solutions for graph optimization.

Learning-based methods can adapt to dynamic network environments.

The survey highlights current challenges and future directions in the field.

Abstract

Existing approaches to solving combinatorial optimization problems on graphs suffer from the need to engineer each problem algorithmically, with practical problems recurring in many instances. The practical side of theoretical computer science, such as computational complexity, then needs to be addressed. Relevant developments in machine learning research on graphs are surveyed for this purpose. We organize and compare the structures involved with learning to solve combinatorial optimization problems, with a special eye on the telecommunications domain and its continuous development of live and research networks.