Dynamics of heuristic optimization algorithms on random graphs.pdfVIP

Dynamics of heuristic optimization algorithms on random graphs Martin Weigt Institute for Theoretical Physics, University of G¨ottingen, Bunsenstr. 9, 37073 G¨ottingen, Germany, E-mail: weigt@theorie.physik.uni-goettingen.de 2 (Dated: February 1, 2008) 0 In this paper, the dynamics of heuristic algorithms for constructing small vertex covers (or inde- 0 pendent sets) of finite-connectivity random graphs is analysed. In every algorithmic step, a vertex 2 is chosen with respect to its vertex degree. This vertex, and some environment of it, is covered and n removed from the graph. This graph reduction process can be described as a Markovian dynamics u in the space of random graphs of arbitrary degree distribution. We discuss some solvable cases, in- J cluding algorithms already analysed using different techniques, and develop approximation schemes 4 for more complicated cases. The approximations are corroborated by numerical simulations. 2 PACS numbers: 89.20.-a, 02.50.-r, 89.20.Ff ] h c I. INTRODUCTION e m - Many questions of practical or scientific interest are based on combinatorial optimization problems whose numerical t solution requires time resources growing exponentially with the system size, or more precisely, with the number of a t binary variables needed to encode the problem. These include examples like planning and scheduling problems s . in various real-world applications, optimization of chip design, cryptografic systems