Text this: Combining Multiple Strategies for Multiarmed Bandit Problems and Asymptotic Optimality