A Batch Sequential Halving Algorithm without Performance Degradation

TL;DR

This paper introduces a batch version of the Sequential Halving algorithm for multi-armed bandits, demonstrating theoretically and empirically that batching does not impair its performance, thus improving computational efficiency without loss of effectiveness.

Contribution

We propose a simple batch adaptation of Sequential Halving and prove it maintains performance, validated through experiments showing robustness in fixed-size batch scenarios.

Findings

Batch Sequential Halving matches the performance of the original algorithm.

Theoretical analysis confirms no degradation due to batching.

Experimental results support the robustness of the batch approach.

Abstract

In this paper, we investigate the problem of pure exploration in the context of multi-armed bandits, with a specific focus on scenarios where arms are pulled in fixed-size batches. Batching has been shown to enhance computational efficiency, but it can potentially lead to a degradation compared to the original sequential algorithm's performance due to delayed feedback and reduced adaptability. We introduce a simple batch version of the Sequential Halving (SH) algorithm (Karnin et al., 2013) and provide theoretical evidence that batching does not degrade the performance of the original algorithm under practical conditions. Furthermore, we empirically validate our claim through experiments, demonstrating the robust nature of the SH algorithm in fixed-size batch settings.