BOIL: Learning Environment Personalized Information

TL;DR

BOIL is a scalable method that uses Pagerank and information maximization to help multi-agent systems extract insights from environments, improving long-term strategies in complex tasks.

Contribution

Introduces BOIL, a novel approach combining Pagerank and information maximization for environment insight extraction in multi-agent systems.

Findings

BOIL outperforms heuristic methods in complex environments.

BOIL improves long-term agent performance in coverage and patrolling tasks.

BOIL effectively guides strategies for stochastic reachability.

Abstract

Navigating complex environments poses challenges for multi-agent systems, requiring efficient extraction of insights from limited information. In this paper, we introduce the Blackbox Oracle Information Learning (BOIL) process, a scalable solution for extracting valuable insights from the environment structure. Leveraging the Pagerank algorithm and common information maximization, BOIL facilitates the extraction of information to guide long-term agent behavior applicable to problems such as coverage, patrolling, and stochastic reachability. Through experiments, we demonstrate the efficacy of BOIL in generating strategy distributions conducive to improved performance over extended time horizons, surpassing heuristic approaches in complex environments.