Making Teams and Influencing Agents: Efficiently Coordinating Decision Trees for Interpretable Multi-Agent Reinforcement Learning

TL;DR

This paper introduces HYDRAVIPER, a decision tree-based interpretable MARL algorithm that balances performance and computational efficiency, enabling safe and verifiable multi-agent decision-making in real-world scenarios.

Contribution

HYDRAVIPER is a novel interpretable MARL method that adaptively manages environment interaction budgets to optimize both performance and efficiency.

Findings

HYDRAVIPER matches state-of-the-art performance with less runtime.

It maintains a Pareto frontier of performance across different interaction budgets.

Experiments demonstrate effectiveness in multi-agent coordination and traffic control.

Abstract

Poor interpretability hinders the practical applicability of multi-agent reinforcement learning (MARL) policies. Deploying interpretable surrogates of uninterpretable policies enhances the safety and verifiability of MARL for real-world applications. However, if these surrogates are to interact directly with the environment within human supervisory frameworks, they must be both performant and computationally efficient. Prior work on interpretable MARL has either sacrificed performance for computational efficiency or computational efficiency for performance. To address this issue, we propose HYDRAVIPER, a decision tree-based interpretable MARL algorithm. HYDRAVIPER coordinates training between agents based on expected team performance, and adaptively allocates budgets for environment interaction to improve computational efficiency. Experiments on standard benchmark environments for multi-agent coordination and traffic signal control show that HYDRAVIPER matches the performance of state-of-the-art methods using a fraction of the runtime, and that it maintains a Pareto frontier of performance for different interaction budgets.