Some Remarks on Replicated Simulated Annealing