# What is Multi-agent RL? Definition and meaning

Reinforcement learning where multiple agents learn together, competing or cooperating, with shared or separate rewards.

## Definition

**Multi-agent RL**: Reinforcement learning where multiple agents learn together, competing or cooperating, with shared or separate rewards.

In short: Several learning agents acting in one environment.

## When it is used

Coordinating machines on a production line, dispatching and traffic-like resource problems.

## Learn more

The full guide to **[Multi-agent RL](/tools/multi-agent-rl)** explains how it works, when to use it and how it compares with the alternatives, with examples. The [glossary](/glossary) lists every term on the site from A to Z.

Previous term: [mTLS](/glossary/mtls) · Next term: [Multi-agent systems](/glossary/multi-agent-systems)
