Goal-Oriented Multi-Agent Reinforcement Learning for Decentralized Agent Teams

TL;DR

This paper introduces a decentralized goal-aware MARL framework enabling autonomous vehicles to communicate selectively, improving coordination, success rates, and scalability in complex, dynamic multi-agent navigation tasks.

Contribution

The paper presents a novel decentralized MARL approach with goal-driven communication, enhancing multi-agent coordination without centralized control.

Findings

Significant improvement in task success rates.

Reduced time-to-goal compared to non-cooperative methods.

Stable performance as the number of agents increases.

Abstract

Connected and autonomous vehicles across land, water, and air must often operate in dynamic, unpredictable environments with limited communication, no centralized control, and partial observability. These real-world constraints pose significant challenges for coordination, particularly when vehicles pursue individual objectives. To address this, we propose a decentralized Multi-Agent Reinforcement Learning (MARL) framework that enables vehicles, acting as agents, to communicate selectively based on local goals and observations. This goal-aware communication strategy allows agents to share only relevant information, enhancing collaboration while respecting visibility limitations. We validate our approach in complex multi-agent navigation tasks featuring obstacles and dynamic agent populations. Results show that our method significantly improves task success rates and reduces time-to-goal compared to non-cooperative baselines. Moreover, task performance remains stable as the number of agents increases, demonstrating scalability. These findings highlight the potential of decentralized, goal-driven MARL to support effective coordination in realistic multi-vehicle systems operating across diverse domains.