Optimal Stopping Problems With A Random Time Horizon