Roman Domination Number in Graph Theory
- Roman domination number is defined as the minimum total weight over all vertex labelings (using 0, 1, and 2) that ensure every unlabeled vertex is adjacent to a vertex labeled 2.
- It refines the classical domination number by modeling strategic resource allocation and offers tight bounds and classifications across various graph families.
- Research explores exact values in paths, cycles, and product graphs, and extends the concept to weighted and k-tuple domination variations with practical implications in network resilience.
A Roman dominating function (RDF) on a graph is a labeling such that every vertex with is adjacent to at least one vertex with . The Roman domination number is the minimum total weight over all RDFs on . This invariant refines the classical domination number by encoding an assignment reminiscent of the strategic allocation of resources under constraints—motivated historically by the “Roman Empire defense” scenario—and admits a sharp interplay with combinatorial graph structure. Roman domination is central in extremal graph theory, product graph analysis, algebraic graph theory, and the study of graph criticality, with explicit parameterizations and classifications in broad graph families.
1. Formal Definition and Fundamental Properties
Let be a finite simple graph. An RDF is a function such that every vertex with has at least one neighbor with . The weight of is where . The Roman domination number is defined as
Fundamental inequalities include: where is the minimum dominating set size (Martinez et al., 2021, Ramezani et al., 2016, Kazemi, 2014, Wang et al., 2011, Cera et al., 27 Dec 2025, Kumar et al., 2024). For any nontrivial graph , . In classic families, the following explicit values hold:
- For , for .
- For paths and cycles , (Kazemi, 2014, Samodivkin, 2017).
In weighted graphs, the weighted Roman domination number generalizes to cost functions , with analogous bounds and explicit formulas for standard families (Cera et al., 27 Dec 2025).
2. Characterizations, Bounds, and Structural Results
RDFs induce natural partitions and structural restrictions:
- Any RDF partitions into .
- In a minimum RDF, the subgraph induced by has maximum degree at most one, there are no edges from to , and forms a vertex cover if the graph has no isolated vertices (Samodivkin, 2017).
- If is edgeless, .
Criticality is studied through vertex-critical graphs (v-critical): is v-critical if and for all . A similar notion applies to edge-critical graphs and Roman saturated graphs, where adding certain edges or removing vertices/edges consistently impacts by exactly one (Martínez-Pérez et al., 2013, Samodivkin, 2017).
3. Exact Values and Extremal Graph Classes
Cycles and Paths
The number is completely classified for paths and cycles: On cycles, criticality matches residue classes modulo 3; is vertex-critical for (Martínez-Pérez et al., 2013).
Roman Domination in Graph Products
For direct (tensor) products : where is the packing number, the total domination number, and the total Roman domination number (Martinez et al., 2021).
For rooted products , only three possible values: with precise structural characterizations for each case depending on the change in upon root removal (Martinez et al., 2021).
Sierpiński Graphs
For the generalized Sierpiński graphs : with explicit evaluations for paths, cycles, and graphs with a universal vertex, as well as tightness in certain cases (Ramezani et al., 2016).
Generalized Petersen Graphs
For , , based on a repeating structure with block period 7 and local accounting arguments (Wang et al., 2011).
Zero-Divisor Graphs and Total Graphs
For commutative ring zero-divisor graphs :
- If is a local ring, .
- If , each a domain, .
- For the total graph , bounds apply (Kumar et al., 2024).
4. Roman Domination Under Graph Operations and Perturbations
The Roman domination number's behavior under graph modifications is systematized:
- Removal of a vertex lowers by 1 if and only if receives label 1 in some optimum RDF. If in every minimum RDF has label 2, (Samodivkin, 2017).
- Addition of edges, or deletion of edges, yields tight bounds or sharply characterized "critical" and "unchanging" Roman domination classes; forests of stars and double-stars provide prototypical examples for each regime (Samodivkin, 2017).
- Structural Venn diagrams relate all six main classes (vertex-reducing, vertex-unchanging, edge-reducing, etc.) with explicit inclusion and intersection properties.
5. Extensions: Weighted and -Tuple Roman Domination
Weighted Roman domination generalizes to graphs with positive vertex weights , seeking a minimum total cost . For weighted graphs,
where is the (weighted) differential (Cera et al., 27 Dec 2025). Exact values are given for , , and cycles.
The Roman -tuple domination number requires each 0-labeled vertex to have at least neighbors with label 2, interpolating between classical and more redundant domination scenarios. For , one recovers the classical Roman domination number (Kazemi, 2014).
6. Applications, Algorithmic Complexity, and Open Problems
- Roman domination is NP-hard to compute even in unweighted graphs; the weighted problem retains this complexity (Cera et al., 27 Dec 2025).
- Explicit constructions and block or recursive methods yield tight values in highly symmetric or product graphs (Wang et al., 2011, Ramezani et al., 2016, Martinez et al., 2021).
- Classification problems include determining all graphs with a given Roman domination number, all trees that are -excellent (for each vertex, there exists a minimum RDF avoiding label 0 at that vertex) (Samodivkin, 2016), and understanding Roman criticality (the effect of specific edge or vertex changes) (Martínez-Pérez et al., 2013).
- For product graphs, a complete classification of which of the three values for occurs, in terms of the so-called "Roman-criticality" of vertices/roots, is an open direction (Martinez et al., 2021).
Roman domination parameters provide a powerful lens for probing redundancy, resilience, and allocation strategies in combinatorial, algebraic, and algorithmic settings, with a mature set of extremal results and open links to more general -domination and weighted invariants.