Optimal Algorithms for Range Searching over Multi-Armed Bandits
This work addresses a stochastic optimization problem for researchers in bandit algorithms and computational geometry, but it is incremental as it adapts a known geometric construct to a bandit setting.
The paper tackles the problem of range searching with stochastic weights in multi-armed bandits, developing sample-efficient algorithms that find near-optimal arms within given intervals with PAC guarantees, and proves these sample complexities are essentially tight.
This paper studies a multi-armed bandit (MAB) version of the range-searching problem. In its basic form, range searching considers as input a set of points (on the real line) and a collection of (real) intervals. Here, with each specified point, we have an associated weight, and the problem objective is to find a maximum-weight point within every given interval. The current work addresses range searching with stochastic weights: each point corresponds to an arm (that admits sample access) and the point's weight is the (unknown) mean of the underlying distribution. In this MAB setup, we develop sample-efficient algorithms that find, with high probability, near-optimal arms within the given intervals, i.e., we obtain PAC (probably approximately correct) guarantees. We also provide an algorithm for a generalization wherein the weight of each point is a multi-dimensional vector. The sample complexities of our algorithms depend, in particular, on the size of the optimal hitting set of the given intervals. Finally, we establish lower bounds proving that the obtained sample complexities are essentially tight. Our results highlight the significance of geometric constructs -- specifically, hitting sets -- in our MAB setting.