Parabellum

Ultra-scalable JaxMARL based warfare simulation engine developed with Armasuisse funding.

Features

• Obstacles and terrain integration
• Rasterized maps
• Blast radii simulation
• Friendly fire mechanics
• Pygame visualization
• JAX-based parallelization

Install

pip install parabellum

Quick Start

import parabellum as pb
from jax import random

terrain = pb.terrain_fn("Thun, Switzerland", 1000)
scenario = pb.make_scenario("Thun", terrain, 10, 10)
env = pb.Parabellum(scenario)

rng, key = random.split(random.PRNGKey(0))
obs, state = env.reset(key)

# Simulation loop
for _ in range(100):
    rng, rng_act, key_step = random.split(key)
    key_act = random.split(rng_act, len(env.agents))
    act = {a: env.action_space(a).sample(k) for a, k in zip(env.agents, key_act)}
    obs, state, reward, done, info = env.step(key_step, act, state)

# Visualize
vis = pb.Visualizer(env, state_sequence)
vis.animate()

Documentation

Full documentation: parabellum.readthedocs.io

Team

• Noah Syrkis
• Timothée Anne
• Supervisor: Sebastian Risi

License

MIT