Sciweavers

JCO
1998

A Greedy Randomized Adaptive Search Procedure for the Feedback Vertex Set Problem

13 years 3 months ago
A Greedy Randomized Adaptive Search Procedure for the Feedback Vertex Set Problem
Abstract. A Greedy Randomized Adaptive Search Procedure (GRASP) is a randomized heuristic that has produced high quality solutions for a wide range of combinatorial optimization problems. The NP-complete Feedback Vertex Set (FVS) Problem is to find the minimum number of vertices that need to be removed from a directed graph so that the resulting graph has no directed cycle. The FVS problem has found applications in many fields, including VLSI design, program verification, and statistical inference. In this paper, we develop a GRASP for the the FVS problem. We describe GRASP construction mechanisms and local search, as well as some efficient problem reduction techniques. We report computational experience on a set of test problems using three variants of GRASP. Key words. Combinatorial optimization, feedback vertex set, local search, GRASP, probabilistic algorithm, computer implementation, heuristics, integer programming
Panos M. Pardalos, Tianbing Qian, Mauricio G. C. R
Added 22 Dec 2010
Updated 22 Dec 2010
Type Journal
Year 1998
Where JCO
Authors Panos M. Pardalos, Tianbing Qian, Mauricio G. C. Resende
Comments (0)