Stigmergy ablation
Method: In-tree honest proxies — not upstream Govcraft LLM meeting-scheduling (48.5% paper benchmark).
Summary
| Bench | Stigmergy | Random / baseline | Notes |
|---|---|---|---|
Chain dispatch (stigmergy_ablation) | 5/5 zones | 0/5 zones | Dependency wake via pheromone deposits |
| Gridworld success rate (8 runs) | 100% | 100% | Both configs reach target; chain dispatch is the discriminant |
Govcraft proxy (govcraft_acceptance) | 41.7% booking rate | 8.3% booking rate | Discrete slot scheduling with PheromoneField |
| Upstream Govcraft | — | — | Requires API keys / Ollama; not bundled in-tree |
Raw JSON (CI reproduce)
{"stigmergy_chain_completed":5,"random_chain_completed":0,"stigmergy_success_rate":1.0,"random_success_rate":1.0}
{"bench":"govcraft_acceptance","mode":"in_tree_honest_equivalent","stigmergy_bookings":200,"random_bookings":40,"stigmergy_rate":0.4166666666666667,"random_rate":0.08333333333333333}
Honest labeling
- In-tree proxy:
benches/govcraft_acceptanceandbenches/stigmergy_ablationusePheromoneField+ scheduler semantics shaped after Govcraft pressure-field-experiment. - Not reproduced: LLM agent meeting scheduling at 48.5% vs 12.6% conversation baseline from the January 2026 paper.
- CI: Both benches run in
./scripts/verify.shon every PR.
Reproduce
cargo run -q -p stigmergy_ablation
cargo run -q -p govcraft_acceptance
bash docs/demo/run-ablation.sh
./scripts/verify.sh