On the Confluence of Directed Graph Reductions Preserving Feedback Vertex Set Minimality

Moussa Abdenbi,Alexandre Blondin Massé,Alain Goupil,Odile Marcotte

DOI: https://doi.org/10.4204/EPTCS.403.6

2024-06-24

Abstract:In graph theory, the minimum directed feedback vertex set (FVS) problem consists in identifying the smallest subsets of vertices in a directed graph whose deletion renders the directed graph acyclic. Although being known as NP-hard since 1972, this problem can be solved in a reasonable time on small instances, or on instances having special combinatorial structure. In this paper we investigate graph reductions preserving all or some minimum FVS and focus on their properties, especially the Church-Rosser property, also called confluence. The Church-Rosser property implies the irrelevance of reduction order, leading to a unique directed graph. The study seeks the largest subset of reductions with the Church-Rosser property and explores the adaptability of reductions to meet this criterion. Addressing these questions is crucial, as it may impact algorithmic implications, allowing for parallelization and speeding up sequential algorithms.

Discrete Mathematics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is about the confluence of graph reductions for the Minimum Directed Feedback Vertex Set (MFVS) problem in directed graphs. Specifically, the authors explore which graph reduction methods can preserve the minimality of MFVS, and whether these reductions satisfy the Church - Rosser property (i.e., confluence), so as to ensure that the reduction order is irrelevant and ultimately obtain a unique reduced graph.

### Problem Background

In graph theory, the minimum directed feedback vertex set problem refers to finding the smallest subset of vertices in a directed graph such that after deleting these vertices, the graph becomes acyclic. Computing the minimum feedback vertex set is an NP - hard problem. To simplify this problem, researchers have proposed various graph reduction methods, which can reduce the size of the graph without changing the minimum feedback vertex set.

### Research Objectives

1. **Explore graph reductions that preserve MFVS**: The authors study which graph reduction methods can simplify the graph while preserving all or part of the minimum feedback vertex sets.
2. **Verify confluence**: Pay special attention to whether these reduction methods satisfy the Church - Rosser property, that is, regardless of the order of reduction, the same reduced graph can be finally obtained.
3. **Algorithm implications**: Explore the impact of these reduction methods on algorithm design, including the potential for parallelization and the possibility of accelerating sequential algorithms.

### Main Contributions

- **Identify confluent reductions**: The authors identify a set of graph reduction methods with confluence, which helps to better understand the properties of graph reduction and its impact on algorithm design.
- **Theoretical verification**: Through strict mathematical proofs, verify the confluence of certain reduction methods, providing a theoretical basis for future algorithm optimization.
- **Future work directions**: Propose further research on the DOME reduction method and explore how to modify it so that it can be included in the set of confluent reductions.

### Conclusions

This research not only enhances our understanding of graph reduction strategies but also provides new ideas for developing more efficient algorithms, especially in terms of parallelization and accelerating sequential algorithms.

### Formulas and Notations

- \( G=(V, E) \) represents a directed graph, where \( V \) is the vertex set and \( E\subseteq V\times V \) is the edge set.
- \( N^+_G(u)=\{s\in V\mid (u, s)\in E\} \) represents the set of successor nodes of vertex \( u \).
- \( N^-_G(u)=\{p\in V\mid (p, u)\in E\} \) represents the set of predecessor nodes of vertex \( u \).
- \( G - u \) represents the graph obtained by deleting vertex \( u \) and all its associated edges from graph \( G \).
- \( G-(u, v) \) represents the graph obtained by deleting edge \((u, v) \) from graph \( G \).
- \( G\circ u \) represents the graph obtained by deleting vertex \( u \) and all its associated edges from graph \( G \), and adding new edges \((p, s) \) for all \( p\in N^-_G(u) \) and \( s\in N^+_G(u) \).

Hopefully, this information will help you better understand the research content of this paper. If you have more questions, feel free to continue asking!