Skip to content

Example 1: Base Simulation (Case 0)

This scenario demonstrates the simplest CosimGym layout: one federation wiring a physics model (spring_mass_damper) to a signal generator (inputs4spring).

Scenario Source

Found at src/scenarios/simple_test.yaml.

Core Takeaways

1. Topology

A single federation federation_1 runs a zmq broker and hosts two base federates:

  • input_federate → runs inputs4spring, publishes force and disturbance.
  • spring_federate → runs spring_mass_damper, subscribes to those signals and publishes position, velocity, acceleration.

Both federates instantiate 2 model instances (n_instances: 2), so two independent springs run in parallel.

2. Model instantiation

model_configs is a single object with an instantiation block (not a list of models):

model_configs:
  instantiation:
    model_name: "spring_mass_damper"
    n_instances: 2
    prefix: "spring"
  parameters:
    mass:      [5, 5]          # one value per instance
    damping:   [2.0, 3.0]
    stiffness: [10.0, 20]
  init_state:
    position: [0.0, 100.0]
    velocity: 0.0
    force:    [10.0, 5.0]

3. Publish / Subscribe wiring

The spring subscribes its force input to the input federate's force publication, per instance. The targets map is keyed by instance index:

subscribes:
  - key: "force"
    type: "double"
    units: "N"
    targets:
      '0': [input_federate.0/force]   # <federate_name>.<instance>/<pub_key>
      '1': [input_federate.1/force]

Execution

Set the scenario in the entry-point script and run it (no --scenario flag exists):

# src/test_script.py
main('simple_test')

conda activate cosim_gym
python src/test_script.py        # or: make run

Once complete, open the dashboard (make run-dashboard) and select the run. You will see the recorded position, velocity, and acceleration traces for each spring instance responding to the driving force.