Optimal Cooperative Multiplayer Learning Bandits with Noisy Rewards and No Communication