Reproducing the ICCD paper

Target paper: ICCD: CRDT-Based Intent Propagation over Contact-Plan DTN.

Versions pinned for the canonical results

Component	Version	SHA256 (short)
`gossamer_threaded_intelligence` wheel	TBD (paper rev)	TBD
`leviathan_engine` wheel	TBD	TBD
`maneuver-map` backend	TBD (git)	TBD
Python	3.11.x	—

Fill this table from the provenance block of the canonical experiment.json; replace TBD values with the actual hashes that shipped with the paper.

Environment setup


python3.11 -m venv .venv && source .venv/bin/activate
pip install -r maneuver-map/backend/requirements.txt
# Or, against a pinned wheel index:
pip install --extra-index-url https://us-central1-python.pkg.dev/arboria-research/python-packages/simple/ \
    gossamer-threaded-intelligence==<pinned> \
    leviathan==<pinned>
export ENGINE_MODE=inprocess   # skip HTTP for deterministic wall-clock
export PYTHONHASHSEED=0

Trial A — ICCD + contact-plan DTN

Canonical body for POST /experiments:


{
  "name": "iccd_contact_plan_inner_heliosphere",
  "steps": 12000,
  "num_agents": 50000,
  "dt": 0.5,
  "generations": 1,
  "output_frequency": 20,
  "environment_bound": 2.25e11,
  "flock_params": {
    "alignment_weight": 1.0,
    "cohesion_weight": 1.0,
    "separation_weight": 1.6,
    "neighbor_radius": 120000.0,
    "separation_distance": 500.0,
    "max_speed": 50.0
  },
  "algorithm": "flocking",
  "algo_params": {
    "iccd": { "initial_aoi": 14400.0, "relay_rotation": true },
    "dtn":  { "contact_density": 0.15, "bandwidth_kbps": 8.0,
              "one_way_delay_sec": 10800.0, "custody_transfer": true },
    "energy": { "soc_wh": 120.0, "radio_j_per_kb": 0.3 },
    "failure": { "relay_attrition_rate_per_hour": 0.002 }
  },
  "seed_base": 42
}

Expected wall-clock on the reference machine (32-core CPU, 256 GB): ~3.5 hours for a full 1e6-agent run; the 50k-agent trial above is ~8 minutes and produces the metrics for Table 1 in the paper.

Expected outputs

After the run completes, the experiment.json at /data/experiments/<exp_id>/experiment.json should contain:


{
  "final_swarm_metrics": {
    "cohesion": "~ expected value, fill from canonical run",
    "alignment": "~ expected value",
    "separation": "~ expected value"
  },
  "paper_metrics": {
    "coverage_ratio": "~ expected",
    "collision_rate_per_agent_step": "~ expected"
  }
}

Fill in the canonical values from the archived run. The canonical experiment.json blob is published alongside the paper’s DOI.

Cross-reference with baselines

Trials B (Periodic Broadcast) and C (Epidemic Flooding) live in the same format; copy the appropriate config body from §4.7 of the paper and submit with the same seed_base=42 so the comparison is apples-to-apples.