Hardness and Parameterized Algorithms on Rainbow Connectivity problem

A path in an edge colored graph is said to be a rainbow path if no two edges on the path have the same color. An edge colored graph is (strongly) rainbow connected if there exists a (geodesic) rainbow path between every pair of vertices. The (strong) rainbow connectivity of a graph G, denoted by (src(G), respectively) rc(G) is the smallest number of colors required to edge color the graph such that G is (strongly) rainbow connected. In this paper we study the rainbow connectivity problem and the strong rainbow connectivity problem from a computational point of view. Our main results can be summarised as below: 1) For every fixed k >= 3, it is NP-Complete to decide whether src(G) <= k even when the graph G is bipartite. 2) For every fixed odd k >= 3, it is NP-Complete to decide whether rc(G) <= k. This resolves one of the open problems posed by Chakraborty et al. (J. Comb. Opt., 2011) where they prove the hardness for the even case. 3) The following problem is fixed parameter tractable: Given a graph G, determine the maximum number of pairs of vertices that can be rainbow connected using two colors. 4) For a directed graph G, it is NP-Complete to decide whether rc(G) <= 2.

💡 Research Summary

The paper studies two closely related graph‑coloring problems: rainbow connectivity (rc) and strong rainbow connectivity (src). A rainbow path is a path whose edges all receive distinct colors; a graph is rainbow (resp. strongly rainbow) connected if every pair of vertices is linked by a rainbow (resp. geodesic rainbow) path. The rc(G) (resp. src(G)) of a graph G is the minimum number of colors needed to achieve rainbow (resp. strong rainbow) connectivity. The authors investigate the computational complexity of deciding whether rc(G) ≤ k or src(G) ≤ k for a given integer k, and they also explore parameterized and approximation aspects of these problems.

Main contributions

1. Strong rainbow connectivity is NP‑complete for any fixed k ≥ 3, even on bipartite graphs.
   The authors introduce the k‑subset strong rainbow connectivity problem, where only a prescribed set P of vertex pairs must be geodesically rainbow‑connected. By a simple reduction from the classic k‑vertex‑coloring problem they show that this subset version is NP‑hard even when the underlying graph is a star. They then transform an instance of the subset problem into a bipartite graph G′ by adding auxiliary vertices (u_i, u′i for each leaf and w{i,j}, w′_{i,j} for each non‑P pair) and a few extra edges. The construction forces any strong rainbow coloring of G′ with k colors to induce a proper vertex coloring of the original graph, establishing NP‑completeness of src(G) ≤ k on bipartite graphs. As a corollary, deciding rc(G) ≤ 3 is also NP‑hard on bipartite graphs.

2. Rainbow connectivity is NP‑complete for every odd k ≥ 3.
   Prior work had shown hardness only for even k. The authors reduce the k‑subset rainbow connectivity problem (already known to be NP‑hard) to the ordinary rc problem for odd k = 2m + 1. For each original vertex v_i they create two chains of length m, connect the ends of the chains, and add two universal vertices x and y that attach to the chain ends. For each non‑P pair they insert cross‑edges between the far ends of the corresponding chains. The design guarantees that any pair in P can be rainbow‑connected only via a short (length‑2) path, while any non‑P pair would need a longer path that inevitably repeats a color if only k colors are allowed. Hence rc(G) ≤ k is NP‑complete for all odd k ≥ 3, completing the dichotomy.

3. Maximum rainbow‑connected pairs with two colors is fixed‑parameter tractable.
   The authors consider the optimization version: given a graph G and a parameter t, decide whether at least t vertex pairs can be rainbow‑connected using only two colors. By viewing the problem as a special case of subset rainbow connectivity, they develop a kernelization that reduces the instance to O(t²) vertices, followed by a bounded‑search‑tree algorithm. The total running time is f(t)·poly(n), proving the problem is FPT with respect to t.

4. Directed graphs: rc(G) ≤ 2 is NP‑complete.
   Extending the notion to digraphs, the authors give a reduction from 3‑SAT. Variables and clauses become vertices; directed edges encode literal occurrences. A two‑coloring of the digraph yields a rainbow path between every ordered pair iff the original formula is satisfiable. Consequently, deciding rc(G) ≤ 2 for directed graphs is NP‑complete.

5. Approximation hardness for strong rainbow connectivity.
   Using the reduction from vertex coloring, the constructed graph G′′ has Θ(n²) vertices, and src(G′′) ≤ k iff the original graph is k‑colorable. If there existed a polynomial‑time algorithm approximating src within a factor n^{1/2‑ε}, it would imply an n^{1‑ε}‑approximation for chromatic number, contradicting known hardness results unless NP = ZPP. Therefore src cannot be approximated within n^{1/2‑ε} unless NP = ZPP.

Technical approach
The paper’s core technique is a series of carefully crafted polynomial‑time reductions that preserve the “rainbow” property while embedding classic hard problems (vertex coloring, 3‑SAT, subset rainbow connectivity) into the rainbow‑connectivity framework. The reductions exploit the fact that in a star or a tree there is a unique simple path between any two leaves, making it easy to force a rainbow condition to correspond exactly to a vertex‑coloring constraint. For the strong version, additional gadgets ensure that any geodesic rainbow path must pass through a central hub, again mirroring vertex colors. For the odd‑k rainbow case, the authors build layered chain gadgets and universal connectors to separate “allowed” short paths from “forbidden” longer ones, thereby encoding the subset requirement. The FPT result relies on kernelization based on the number of desired rainbow pairs, while the directed‑graph hardness uses a standard SAT‑to‑graph encoding with a two‑color restriction.

Related work and context
The paper situates its contributions within a growing literature on rainbow connectivity. Chartrand et al. introduced the concept; Caro et al., Krivelevich & Yuster, and Chandran et al. provided bounds and probabilistic analyses. Chakraborty et al. proved NP‑hardness for rc = 2 and for even k, leaving the odd‑k case open. This work resolves that open problem, extends hardness to bipartite graphs, and adds parameterized and approximation perspectives that were previously missing.

Implications and future directions
The results show that even very restricted versions of rainbow connectivity (bipartite graphs, two colors, directed graphs) remain computationally intractable, suggesting that practical applications will need heuristic or approximation methods. The FPT algorithm for the two‑color case opens the door to further parameterized studies (e.g., parameterizing by treewidth, feedback vertex set size, or the number of colors). The approximation lower bound for src hints that any efficient approximation scheme must either accept large factors or exploit additional structural graph properties. Future work could explore tighter bounds for special graph classes (planar, bounded degree), develop exact exponential‑time algorithms, or investigate the interplay between rainbow connectivity and other connectivity notions such as edge‑disjoint paths.

Overall, the paper delivers a comprehensive complexity landscape for rainbow and strong rainbow connectivity, resolves a standing open problem, and introduces new algorithmic avenues through parameterized analysis.