Decision-focused Graph Neural Networks for Combinatorial Optimization

Yang Liu,Chuan Zhou,Peng Zhang,Shirui Pan,Zhao Li,Hongyang Chen

2024-06-10

Abstract:In recent years, there has been notable interest in investigating combinatorial optimization (CO) problems by neural-based framework. An emerging strategy to tackle these challenging problems involves the adoption of graph neural networks (GNNs) as an alternative to traditional algorithms, a subject that has attracted considerable attention. Despite the growing popularity of GNNs and traditional algorithm solvers in the realm of CO, there is limited research on their integrated use and the correlation between them within an end-to-end framework. The primary focus of our work is to formulate a more efficient and precise framework for CO by employing decision-focused learning on graphs. Additionally, we introduce a decision-focused framework that utilizes GNNs to address CO problems with auxiliary support. To realize an end-to-end approach, we have designed two cascaded modules: (a) an unsupervised trained graph predictive model, and (b) a solver for quadratic binary unconstrained optimization. Empirical evaluations are conducted on various classical tasks, including maximum cut, maximum independent set, and minimum vertex cover. The experimental results on classical CO problems (i.e. MaxCut, MIS, and MVC) demonstrate the superiority of our method over both the standalone GNN approach and classical methods.

Machine Learning,Artificial Intelligence

What problem does this paper attempt to address?

The paper aims to address Combinatorial Optimization Problems (COPs), particularly focusing on classic combinatorial optimization problems such as MaxCut, Maximum Independent Set (MIS), and Minimum Vertex Cover (MVC). The paper proposes a method based on the Decision-Focused Learning (DFL) framework, which combines Graph Neural Networks (GNNs) with traditional optimization algorithms to improve the efficiency and accuracy of solving these complex problems.

Specifically, the authors propose a new framework called G-DFL4CO, which consists of two modules:
1. **Unsupervised Training Graph Prediction Model**: Used to predict the entire graph structure.
2. **Optimizer based on Quadratic Unconstrained Binary Optimization (QUBO)**: Used to solve specific combinatorial optimization problems.

By combining these two modules, the framework can efficiently solve combinatorial optimization problems in an end-to-end setting and demonstrates its advantages over using GNNs or traditional methods alone in experiments. The experimental results show that G-DFL4CO achieves excellent performance on multiple benchmark datasets, particularly excelling in the MaxCut problem. Additionally, the framework also demonstrates faster runtime compared to other methods.