A Stochastic limit approach to the SAT problem