Modelling the dynamics of stochastic local search on k-SAT

摘要

A new analytical tool is presented to provide a better understanding of the search space of k-SAT. This tool, termed the local value distribution, describes the probability of finding assignments of any value q' in the neighbourhood of assignments of value q. The local value distribution is then used to define a Markov model to model the dynamics of a corresponding stochastic local search algorithm for k-SAT. The model is evaluated by comparing the predicted algorithm dynamics to experimental results. In most cases the fit of the model to the experimental results is very good, but limitations are also recognised.