Planet Wars RTS
Multi-Agent AI Competition
Upcoming Competitions 2026
AAMAS 2026
International Conference on Autonomous Agents and Multiagent Systems
When: May 25-29, 2026
Where: Paphos, Cyprus
Website: cyprusconferences.org/aamas2026
Competition details and submission guidelines coming soon!
IEEE WCCI 2026
IEEE World Congress on Computational Intelligence
When: June 21-26, 2026
Where: Maastricht, Netherlands
Website: attend.ieee.org/wcci-2026
Competition details and submission guidelines coming soon!
About the Competition
Planet Wars is a real-time strategy (RTS) game where AI agents compete to gain control of planets and destroy enemy units. The challenge is to develop agents that can play well across a wide range of game parameters and against diverse opponent strategies.
Key Features
Flexible Game Modes
Compete in both fully observable and partially observable game variants
Real-Time Strategy
Handle simultaneous moves and unpredictable opponent actions
Varied Parameters
Test your agents across different planet counts, growth rates, and transit speeds
Multiple Languages
Develop agents in Kotlin, Java, Python, or any language with WebSocket support
What Makes This Challenging?
- Generalization: Agents must adapt to varying game parameters
- Real-Time Decisions: Limited time per decision (typically 50ms)
- Partial Observability: Handle uncertainty about hidden information
- Strategic Depth: Balance expansion, defense, and attack strategies
- Opponent Modeling: Adapt to different opponent playstyles
Past Competitions
GECCO 2025
Genetic and Evolutionary Computation Conference
Winner: TeamTitans (Frankfurt University of Applied Sciences)
Competed in both fully observable and partially observable modes with 11 entries across varied game parameters.
IEEE CoG 2025
IEEE Conference on Games
Winners:
- TeamTitansV3
- NanaBoshi
- GalacticArmada
Focused on fully observable mode with 33 agent submissions.
How to Participate
1. Develop Your Agent
Create an AI agent that implements the PlanetWarsAgent interface. You can use:
- Kotlin/Java: Full access to forward model for planning algorithms (MCTS, etc.)
- Python: Ideal for neural network-based agents (PyTorch, TensorFlow)
- Other Languages: Any language with WebSocket and JSON support
2. Test Locally
Run games with the visual interface or headless mode to develop and debug your strategy.
3. Containerize
Package your agent in a Docker/Podman container that exposes a WebSocket server on port 8080.
4. Submit
Create a GitHub issue in the submissions repository with your containerized agent details.
Get Started: Check out the detailed Submission Instructions for complete guidance on developing and submitting your agent.
Resources
- Main Repository - Source code and examples
- Submission Guide - How to submit your agent
- Submissions Repository - Submit and track entries
- Live Gameplay Demo - Watch AI agents compete
- Python Example - Complete Python agent with neural network
Contact & Links
For questions, suggestions, or collaboration opportunities:
- Open an issue on GitHub
- Check the repository README for updates
- Follow competition-specific pages for detailed guidelines and deadlines