Resource Allocation in Multi-armed Bandit Exploration: Overcoming Sublinear Scaling with Adaptive Parallelism

TL;DR

This paper investigates resource allocation strategies in multi-armed bandit problems with divisible resources, balancing exploration speed and throughput to optimize arm identification under fixed confidence and deadline constraints.

Contribution

It introduces novel algorithms that optimize resource distribution in parallel bandit exploration, accounting for nonlinear scaling effects and providing theoretical bounds and practical performance evaluations.

Findings

Algorithms achieve near-optimal exploration times.

Trade-offs between resource allocation and throughput are effectively managed.

Simulation results confirm improved performance over baselines.

Abstract

We study exploration in stochastic multi-armed bandits when we have access to a divisible resource that can be allocated in varying amounts to arm pulls. We focus in particular on the allocation of distributed computing resources, where we may obtain results faster by allocating more resources per pull, but might have reduced throughput due to nonlinear scaling. For example, in simulation-based scientific studies, an expensive simulation can be sped up by running it on multiple cores. This speed-up however, is partly offset by the communication among cores, which results in lower throughput than if fewer cores were allocated per trial to run more trials in parallel. In this paper, we explore these trade-offs in two settings. First, in a fixed confidence setting, we need to find the best arm with a given target success probability as quickly as possible. We propose an algorithm which trades off between information accumulation and throughput and show that the time taken can be upper bounded by the solution of a dynamic program whose inputs are the gaps between the sub-optimal and optimal arms. We also prove a matching hardness result. Second, we present an algorithm for a fixed deadline setting, where we are given a time deadline and need to maximize the probability of finding the best arm. We corroborate our theoretical insights with simulation experiments that show that the algorithms consistently match or outperform baseline algorithms on a variety of problem instances.