Incentivized Learning in Principal-Agent Bandit Games
This addresses incentive design in applications like healthcare or taxation, extending bandit problems to include learning aspects often overlooked in mechanism design.
The paper tackles the problem of a principal learning to offer incentives to an agent with misaligned objectives in repeated bandit games, achieving nearly optimal regret bounds in multi-armed and linear contextual settings.
This work considers a repeated principal-agent bandit game, where the principal can only interact with her environment through the agent. The principal and the agent have misaligned objectives and the choice of action is only left to the agent. However, the principal can influence the agent's decisions by offering incentives which add up to his rewards. The principal aims to iteratively learn an incentive policy to maximize her own total utility. This framework extends usual bandit problems and is motivated by several practical applications, such as healthcare or ecological taxation, where traditionally used mechanism design theories often overlook the learning aspect of the problem. We present nearly optimal (with respect to a horizon $T$) learning algorithms for the principal's regret in both multi-armed and linear contextual settings. Finally, we support our theoretical guarantees through numerical experiments.