Author

Jacob Adkins

Date of Award

2022

Document Type

Thesis

Degree Name

Bachelors

Department

Natural Sciences

First Advisor

Gillman, David

Area of Concentration

Applied Mathematics and Computer Science

Abstract

This thesis provides a friendly introduction to the field of multi-armed bandits as a framework to consider the exploration-exploitation dilemma. We summarize recent literature that combines the established theory of bandits with recent ideas involving Bayesian-persuasion from algorithmic economics to consider a three-pronged exploration-exploitation-persuasion trade-off. We include initial empirical results concerning the Bayesian regret bounds for a recent incentive compatible bandit algorithm. We also provide an outline of ideas for future study in the area of incentivized multiagent bandit exploration.

Share

COinS