Optimal Monetary Policy with Imperfect Common Knowledge