Run The Ecosystem Simulation

The ecosystem simulation is a BEAM-native blueprint that stresses MirrorNeuron with stateful regions, animal populations, cross-region messages, and a final leaderboard.

Path:

mn-blueprints/ecosystem_summation

What It Simulates

Animals have compact DNA traits:

• metabolism
• forage
• breed
• aggression
• move
• longevity

Regions have local resource profiles. Animals compete for food, age, die, reproduce, mutate, and migrate between neighboring regions. The final output reports high-performing DNA profiles.

Why This Example Is BEAM-Native

This blueprint keeps simulation state in BEAM agent state instead of launching one sandbox per animal.

ingress router
      |
      v
world agent
      |
      v
region agents <--> region agents
      |
      v
collector
      |
      v
summarizer

This design keeps:

• clear state ownership
• cheap message passing
• bounded runtime processes
• lower Redis and sandbox overhead than per-entity workers

Run The Checked-In Blueprint

mn blueprint run ecosystem_simulation

Expected output:

Job submitted successfully

Inspect the job:

mn job list
mn job status <job_id>
mn job monitor <job_id>

When To Use This Example

Use it after simpler examples when you want to test:

• many messages
• stateful agents
• aggregation
• cluster scheduling
• runtime recovery under a larger workload

For first-time setup, use Quickstart instead.

Related Pages

• Examples
• Runtime Architecture
• Cluster Guide
• Reliability Guide

unlink(content/docs/md-legacy/simulation_example.md)