On the Optimal Amount of Experimentation in Sequential Decision Problems

TL;DR

This paper establishes a precise bound on the optimal experimentation level in sequential decision-making, demonstrating its relevance through a bound on the cut-off in a one-arm bandit problem.

Contribution

It introduces a tight bound on experimentation in optimal strategies and applies it to a specific bandit problem scenario.

Findings

Derived a tight bound on experimentation in sequential decisions

Applied the bound to determine the cut-off in a one-arm bandit problem

Demonstrated the practical relevance of the theoretical result

Abstract

We provide a tight bound on the amount of experimentation under the optimal strategy in sequential decision problems. We show the applicability of the result by providing a bound on the cut-off in a one-arm bandit problem.