P-complete

Minimum Feedback Vertex Set (MFVS)

Category: Graph Theory

Input

A directed graph $G = (V, E)$ that is cyclically reducible (defined below) and a designated vertex $v$.

Problem

Is $v$ contained in the minimum feedback set of $G$ that is computed by the algorithm given by Wang, Lloyd, and Soffa [1]?

Definitions

A vertex $z$ of $G$ is deadlocked if there is a directed path in $G$ from $z$ to a vertex $y$ that lies on a directed cycle. The associated graph of vertex x with respect to G, $A(G,x)$, consists of vertex $x$ and all vertices of $G$ that are not deadlocked if $x$ is removed from $G$. A directed graph is cyclically reducible if and only if there exists a sequence of vertices $(y_1, \ldots, y_k)$ such that each of the graphs $A(G_{i-1},y_i)$ is cyclic, where $G_0 = G$ and $G_i = G_{i-1} - A(G_i, y_i)$, for $1 \leq i \leq k$.

A set is called a feedback vertex set if it contains at least one vertex from every cycle (Karp [2]). It is minimum if no other feedback vertex set has fewer elements.

Status

$\P$-complete (Bovet, De Agostino, and Petreschi [3]).

Remarks

Wang, Lloyd, and Soffa gave a polynomial time algorithm for computing feedback sets in cyclically reducible graphs [1]. Thus, MFVS is in $\P$. The reduction to show hardness is from MCVP.

The question of whether a general graph has a feedback vertex set of size $k$ is $\NP$-complete (Karp [2]). Bovet, De Agostino, and Petreschi give an algorithm for finding a minimum feedback set in cyclically reducible graphs that requires $O(k(\log n)^2)$ time and $O(n^4)$ processors on a CREW-PRAM, where $k$ denotes the size of the minimum feedback set. Greenlaw proved that a related problem, the Lexicographically First Maximal Acyclic Subgraph Problem, is $\P$-complete [4][5].

Ramachandran proved that finding a minimum weight feedback vertex set in a reducible flow graph with arbitrary weights is $\P$-complete [6][7]. She also proved the following four problems are $\NC$ equivalent: finding a minimum feedback arc set in an unweighted reducible flow graph, finding a minimum weight feedback arc set in a reducible flow graph with unary weights on the arcs, finding a minimum weight feedback vertex set in a reducible flow graph with unary weights on the vertices, and finding a minimum cut in a flow network with unary capacities [6][7].

References

  1. C.-C. Wang, E. L. Lloyd, and M. L. Soffa (1985). Feedback vertex sets and cyclically reducible graphs.
  2. R. M. Karp (1972). Reducibility among combinatorial problems.
  3. D. P. Bovet, S. De Agostino, and R. Petreschi (1988). Parallelism and the feedback vertex set problem.
  4. R. Greenlaw (1992). The parallel complexity of approximation algorithms for the maximum acyclic subgraph problem.
  5. R. Greenlaw (1992). Towards understanding the effective parallelization of sequential algorithms.
  6. V. Ramachandran (1988). Fast and processor-efficient parallel algorithms for reducible flow graphs.
  7. V. Ramachandran (1990). Fast parallel algorithms for reducible flow graphs.
  8. Greenlaw, R., Hoover, H. J., and Ruzzo, W. L. (1995). Limits to Parallel Computation: P-Completeness Theory. Oxford University Press. Problem A.2.17.