1 problem

Consider a $K$-armed bandit model. For arm $k$, let $N_{k,T}$ be the number of pulls by time $T$, let $\bar\mu_{k,T}$ be its empirical mean reward, and let $I(\cdot)$ be the index…