What happens in nature?
In Tero and colleagues' experiment, Physarum formed transport networks between food locations. A comparison with Tokyo rail infrastructure examined trade-offs among network cost, transport efficiency and fault tolerance.
From mechanism to protocol
- Initial communication network
- Verified contribution per connection
- Update weight · preserve a fallback path
| In biology | A counterpart to test in your agents |
|---|---|
| Flow-carrying tube | Communication edge with recorded output consumption |
| Network cost | Message tokens, latency and routing overhead |
| Adaptation | Edge-weight updates from verified experiment results |
| Alternative route | Fallback handoff for critical information |
How can you use it in your swarm?
Update communication paths using measured contribution and monitor single-link dependence.
Freeze the initial graph
Record a fixed graph for initial evaluation. Define each edge's information payload and cost unit.
Measure contribution
Link each handoff to information actually used and the final test outcome. Message count is not success.
Update under control
Update only between evaluation rounds. Change few edges per round so the effect remains traceable.
Test the fallback
Remove one strong connection and rerun the task. Report savings alongside quality loss after interruption.
Start with an experiment
Compare fixed and utility-updated graphs on the same task sequence; keep final evaluation questions outside the update loop.
- What happens if you remove the mechanism?
- Remove fallback links, then make one specialist unavailable. Show cost savings and recovery loss together.
- Primary failure risk
- Rewarding frequently used routes can create an echo chamber. Monitor access to rare expertise and distribution shifts.
Evidence and related studies
Read biological evidence and agent research separately. The engineering interpretations below are SWI synthesis.
Rules for biologically inspired adaptive network design
Physarum networks achieved comparable trade-offs in efficiency, cost and fault tolerance in a Tokyo rail-network example.
What can I use? Interpretation, limits and provenance
SWI engineering interpretation
Retain communication edges for utility relative to cost, not frequency alone.
Limitation
Physarum is a multinucleate single-celled organism, not a colony of independent LLM agents.
Source record
Atsushi Tero et al.
Publication: 2010-01-22
Abstract review · Checked: 2026-09-06
AntNet: Distributed Stigmergetic Control for Communications Networks
Distributed mobile agents adapt routing tables through indirect, asynchronous information exchange mediated by the network.
What can I use? Interpretation, limits and provenance
SWI engineering interpretation
Update task routing from observed results rather than fixed role titles.
Limitation
Network packet costs are measurable; LLM correctness requires an additional external evaluator.
Source record
Gianni Di Caro, Marco Dorigo
Publication: 1998-12-01
Abstract review · Checked: 2026-09-06
SwarmSys: Decentralized Swarm-Inspired Agents for Scalable and Adaptive Reasoning
SwarmSys combines explorer, worker and validator roles with adaptive matching and pheromone-inspired reinforcement.
What can I use? Interpretation, limits and provenance
SWI engineering interpretation
Tie trail strength to verified task outcomes rather than an agent's self-reported confidence.
Limitation
Preprint results belong to the authors' experiments; this app has not reproduced that performance.
Source record
Ruohao Li et al.
Publication: 2025-10-11
Abstract review · Checked: 2026-09-06
Evolve as a Team: Collaborative Self-Evolution for LLM-based Multi-Agent Systems
Meta-Team preserves execution context and uses post-task communication to improve agent behavior, coordination and team organization.
What can I use? Interpretation, limits and provenance
SWI engineering interpretation
Turn retrospectives into small protocol changes to evaluate in the next version.
Limitation
A 2026 preprint; distinguish learning from execution from adaptation to test data.
Source record
Zhezheng Hao et al.
Publication: 2026-05-28
Abstract review · Checked: 2026-09-06