Hitting all Maximal Independent Sets of a Bipartite Graph

We prove that given a bipartite graph G with vertex set V and an integer k, deciding whether there exists a subset of V of size k hitting all maximal independent sets of G is complete for the class Sigma_2^P.

💡 Research Summary

The paper investigates the decision problem “Given a bipartite graph G=(V,E) and an integer k, does there exist a vertex subset X⊆V of size k that intersects every maximal independent set of G?” The authors prove that this problem is Σ₂^P‑complete, i.e., complete for the second level of the polynomial‑time hierarchy. The contribution consists of two main parts: (1) showing that the problem belongs to Σ₂^P, and (2) establishing Σ₂^P‑hardness via a polynomial‑time reduction from a known Σ₂^P‑complete problem, namely ∃∀‑3‑SAT (the quantified Boolean formula with an existential block followed by a universal block).

Membership in Σ₂^P.
The authors formalize the problem as the logical statement
∃X⊆V, |X|=k ∀M∈MIS(G) X∩M≠∅,
where MIS(G) denotes the family of all maximal independent sets of G. This is an ∃∀‑formula, which is precisely the shape of Σ₂^P. To verify a candidate set X, one must check that no maximal independent set of G is disjoint from X. This can be done by an NP‑oracle: given X, ask whether G−X contains a maximal independent set (equivalently, whether there exists an independent set that cannot be extended inside G−X). If the oracle answers “yes,” then X fails; if it answers “no,” then X is a valid hitting set. Because the verification uses a single NP query, the overall decision procedure lies in NP^NP = Σ₂^P.

Σ₂^P‑hardness.
The hardness proof is the technical core. The authors start from an instance of ∃∀‑3‑SAT: a formula Φ = ∃X ∀Y ψ(X,Y) where ψ is a 3‑CNF. They construct, in polynomial time, a bipartite graph G_Φ and an integer k such that Φ is true iff G_Φ has a hitting set of size k for all its maximal independent sets.

The construction proceeds as follows:

1. Variable gadgets. For each existential variable x_i a pair of vertices (x_i, ¬x_i) is placed on one side of the bipartition; similarly, each universal variable y_j is represented by a pair (y_j, ¬y_j) on the opposite side. Edges are added between complementary literals to enforce that any independent set can contain at most one literal from each pair.

2. Clause gadgets. For each clause C = (ℓ₁ ∨ ℓ₂ ∨ ℓ₃) a clause vertex c_C is introduced on the side opposite to the literals. It is connected to the three literal vertices that appear negated in the clause. The bipartite nature guarantees that any independent set that contains c_C cannot contain any of those three literals, and vice‑versa.

3. Maximality enforcement. Additional “selector” vertices are added to force every maximal independent set to be of a specific form: either it picks a literal from each variable pair (thus encoding a truth assignment) or it picks the clause vertex, which is only possible when the corresponding clause is unsatisfied by the chosen literals.

4. Size parameter k. The integer k is set to the number of existential variables plus a small constant that accounts for the selector vertices. Intuitively, a hitting set of size k must choose exactly one literal for each existential variable and possibly a few auxiliary vertices; it cannot afford to pick any clause vertex because that would exceed the budget.

The crucial observation is that a maximal independent set that avoids the hitting set corresponds precisely to a universal assignment Y that falsifies ψ under the existential assignment encoded by the hitting set. Consequently, if the existential player has a winning strategy in the original QBF (i.e., Φ is true), then there exists a size‑k hitting set for G_Φ; otherwise, no such hitting set exists. This establishes a many‑one reduction from ∃∀‑3‑SAT to the hitting‑set problem, proving Σ₂^P‑hardness.

Implications and extensions.
By pinning the problem at the second level of the polynomial hierarchy, the paper shows that it is unlikely to admit a polynomial‑time algorithm unless the hierarchy collapses. Moreover, the result distinguishes this hitting‑set problem from classic NP‑complete graph problems such as Vertex Cover or Maximum Independent Set, which lie at the first level. The authors discuss how the same technique can be adapted to other “hitting‑all‑maximal‑structures” problems, for example, hitting all maximal cliques or all minimal dominating sets, suggesting a broader class of Σ₂^P‑complete graph problems.

Finally, the paper outlines possible research directions: (i) investigating fixed‑parameter tractability with respect to k, (ii) designing approximation algorithms or heuristics for practical instances, and (iii) exploring restricted graph families (e.g., trees, bounded‑degree bipartite graphs) where the problem might become easier.

In summary, the authors provide a rigorous complexity classification for the problem of selecting a size‑k vertex set that intersects every maximal independent set in a bipartite graph, proving it to be Σ₂^P‑complete through a careful reduction from quantified Boolean formulas and a clear membership argument. This work deepens our understanding of the interplay between graph‑theoretic hitting problems and higher‑order complexity classes.