Boosting Combinatorial Search Through Randomization

Abstract

Unpredictability in the running time of complete search procedures can often be explained by the phenomenon of "heavy-tailed cost distributions", meaning that at any time during the experiment there is a non-negligible probability of hitting a problem that requires exponentially more time to solve than any that has been encountered before (Gomes et al. 1998a). We present a general method for introducing controlled randomization into complete search algorithms. The "boosted" search methods provably eliminate heavy-tails to the right of the median. Furthermore, they can take advantage of heavy-tails to the left of the median (that is, a nonnegligible chance of very short runs) to dramatically shorten the solution time. We demonstrate speedups of several orders of magnitude for state-of-the-art complete search procedures running on hard, real-world problems. Introduction The time required by complete search methods to solve similar combinatorial problems can be surprisingl...