Physics – Condensed Matter – Disordered Systems and Neural Networks
Scientific paper
2004-12-17
Physics
Condensed Matter
Disordered Systems and Neural Networks
long version of paper in CP 2004
Scientific paper
Many backtracking algorithms exhibit heavy-tailed distributions, in which their running time is often much longer than their median. We analyze the behavior of two natural variants of the Davis-Putnam-Logemann-Loveland (DPLL) algorithm for Graph 3-Coloring on sparse random graphs G(n,p=c/n). Let P_c(b) be the probability that DPLL backtracks b times. First, we calculate analytically the probability P_c(0) that these algorithms find a 3-coloring with no backtracking at all, and show that it goes to zero faster than any analytic function as c \to c^* = 3.847... Then we show that even in the ``easy'' phase 1 < c < c^* where P_c(0) > 0, including just above the emergence of the giant component, the expected number of backtracks is exponentially large with positive probability. To our knowledge this is the first rigorous proof that the running time of a natural backtracking algorithm has a heavy tail for graph coloring. Moreover, our results show that these algorithms take exponential time, not just below the 3-colorability threshold, but just above the degree c=1 at which the giant component first appears. In addition, we give experimental evidence and heuristic arguments that this tail takes the form P_c(b) ~ b^{-1} up to an exponential cutoff.
Jia Haixia
Moore Cristopher
No associations
LandOfFree
How much backtracking does it take to color random graphs? Rigorous results on heavy tails does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with How much backtracking does it take to color random graphs? Rigorous results on heavy tails, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and How much backtracking does it take to color random graphs? Rigorous results on heavy tails will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-159117