Self-stabilizing algorithms for Connected Vertex Cover and Clique decomposition problems

In many wireless networks, there is no fixed physical backbone nor centralized network management. The nodes of such a network have to self-organize in order to maintain a virtual backbone used to route messages. Moreover, any node of the network can be a priori at the origin of a malicious attack. Thus, in one hand the backbone must be fault-tolerant and in other hand it can be useful to monitor all network communications to identify an attack as soon as possible. We are interested in the minimum \emph{Connected Vertex Cover} problem, a generalization of the classical minimum Vertex Cover problem, which allows to obtain a connected backbone. Recently, Delbot et al.~\cite{DelbotLP13} proposed a new centralized algorithm with a constant approximation ratio of $2$ for this problem. In this paper, we propose a distributed and self-stabilizing version of their algorithm with the same approximation guarantee. To the best knowledge of the authors, it is the first distributed and fault-tolerant algorithm for this problem. The approach followed to solve the considered problem is based on the construction of a connected minimal clique partition. Therefore, we also design the first distributed self-stabilizing algorithm for this problem, which is of independent interest.

💡 Research Summary

The paper addresses the problem of constructing a connected virtual backbone in wireless ad‑hoc networks that simultaneously supports efficient routing and comprehensive monitoring of all communications. Classical solutions based on Connected Dominating Sets (CDS) provide connectivity but do not guarantee that every edge is incident to a backbone node, which is required for security monitoring. Conversely, a Vertex Cover (VC) monitors all edges but may be disconnected, making it unsuitable as a routing backbone. The authors focus on the Minimum Connected Vertex Cover (CVC) problem, which seeks a smallest vertex set that both covers all edges and induces a connected subgraph.

The main contribution is a fully distributed, self‑stabilizing algorithm that achieves a 2‑approximation for CVC, matching the best known centralized bound, while operating under the weakest possible daemon assumptions (unfair, distributed). The algorithm is built on a novel intermediate structure: a Connected Minimal Clique Partition (CMCP). A CMCP is a partition of the graph into maximal cliques such that (i) no two cliques can be merged into a larger clique (minimality) and (ii) the union of all non‑trivial cliques (size ≥ 2) forms a connected subgraph.

System Model
The network is modeled as an undirected, simple, connected graph G = (V,E). Each node has a unique identifier and knows its neighbor list. Computation follows the local shared‑memory model: a node can read its own variables and those of its neighbors, write only to its own variables, and execute actions defined by guarded commands with a fixed priority order. The daemon is unfair (it may indefinitely postpone enabled actions unless they are the only enabled ones) but distributed (it always selects at least one enabled node each step).

Algorithm Overview

1. CMCP Construction

   • Clique Leader Election: Each node compares its identifier with those of its neighbors. If no neighbor has a smaller ID, the node declares itself a leader. This deterministic rule guarantees a unique leader per maximal clique and prevents overlapping clique assignments.
   • Clique Formation: A leader gathers all neighbors that are mutually adjacent (i.e., form a complete subgraph with the leader) and declares them members of its clique. The leader broadcasts the clique identifier to its members.
   • Connectivity Maintenance: Leaders exchange information about neighboring cliques through shared edges (cross‑edges). If a cross‑edge exists between two cliques, the leaders merge their cliques into a single connected component. This merging process repeats until no cross‑edge connects distinct non‑trivial cliques, yielding a CMCP where all non‑trivial cliques are mutually connected.

   The CMCP algorithm converges in O(Diam) asynchronous rounds, where Diam is the graph diameter, because each round propagates leader information at most one hop farther. The number of merging steps is bounded by the number of cliques, guaranteeing termination under the unfair daemon.

2. CVC Extraction

   • Intra‑Clique Coverage: All vertices inside each non‑trivial clique are added to the vertex cover. Since a clique is a complete subgraph, this trivially covers every edge internal to the clique.
   • Inter‑Clique Coverage: For each edge that connects two different cliques, at least one endpoint belongs to a clique leader (or a designated representative). The algorithm adds only the leaders of the cliques to the cover, ensuring that every inter‑clique edge is also covered.
   • Connectivity: Because the CMCP guarantees that the set of non‑trivial cliques is connected, and because leaders are part of those cliques, the induced subgraph on the selected vertices is connected.

   The resulting vertex set is a connected vertex cover. By construction, the size of the cover is at most twice the size of an optimal CVC, yielding a 2‑approximation.

Correctness and Complexity
Safety: In any terminal configuration, the CMCP satisfies minimality and connectivity by definition; the derived CVC covers all edges and induces a connected subgraph.
Convergence: Under the unfair distributed daemon, any node that is not yet a leader will eventually see a neighbor with a smaller ID become a leader, after which it either joins that leader’s clique or becomes a leader itself if no smaller ID exists. The merging phase reduces the number of distinct cliques monotonically, guaranteeing termination after a finite number of rounds.
Time Complexity: The CMCP phase requires O(Diam) rounds; the CVC phase is a constant‑time transformation. The total number of atomic steps is O(m), where m is the number of edges, because each edge is examined only a constant number of times (during leader election, clique formation, and cross‑edge detection).

Related Work
The paper surveys prior self‑stabilizing solutions for graph decompositions such as maximal matching, star and triangle partitions, and clustering. While many of these achieve constant‑factor approximations for their respective problems, none address the simultaneous requirements of connectivity and edge coverage inherent to CVC. Existing centralized algorithms for CVC (e.g., the 2‑approximation by Delbot et al.) rely on global knowledge and cannot tolerate transient faults. The presented work is the first to provide a distributed, fault‑tolerant (self‑stabilizing) solution with the same approximation guarantee.

Limitations and Future Directions
The approach assumes unique node identifiers and a static topology; handling anonymous networks or dynamic joins/leaves would require additional mechanisms. Cases where the graph contains only trivial cliques (e.g., trees) need special handling because the CMCP definition focuses on cliques of size ≥ 2. Future research may explore extensions to asynchronous message‑passing models, dynamic topology adaptation, and empirical evaluation on realistic wireless sensor network deployments to assess message overhead and convergence speed.

Conclusion
By introducing a self‑stabilizing construction of a Connected Minimal Clique Partition and leveraging it to extract a Connected Vertex Cover, the authors deliver the first distributed, fault‑tolerant algorithm for the Minimum Connected Vertex Cover problem. The algorithm operates under minimal daemon assumptions, converges in a number of rounds proportional to the network diameter, and guarantees a 2‑approximation of the optimal solution. This work bridges the gap between routing backbone formation and network‑wide security monitoring, offering a robust foundation for future self‑organizing wireless network protocols.