A computational programme with stated boundaries
A distributed algorithm specifies how separate agents update their states using limited information. This is an engineering object that can be analysed independently of any animal. AESRD proposes to use such objects to sharpen questions about coordination, while keeping implementation claims separate from biological interpretation.
No current AESRD evidence establishes that eusocial shrimp run a consensus protocol, elect a leader, or maintain a shared ledger.
A tractable agreement problem
Consider a fixed finite undirected graph with symmetric, nonnegative weights aᵢⱼ = aⱼᵢ ≥ 0, connected through its positive-weight edges. Let L = D − A, where Dᵢᵢ is the sum of row i of A. The continuous-time rule dx/dt = −Lx drives scalar states toward their initial average. This restricted example offers a clean way to discuss topology and convergence. Directed links, delays, or changing graphs require additional conditions.
The proposed programme would show the assumptions beside every equation. A guarantee proved for one system should not travel unnoticed into another.
Events, delays, and failures
Realistic computational questions include missing messages, limited memory, uncertain timing, and disconnected agents. Lamport’s work on event ordering is a foundational reminder that global chronology is itself a modelling choice in distributed systems.
Demonstrations would compare synchronous updates with asynchronous schedules and record when a result depends on the choice. Failure would be treated as useful information about the model, rather than hidden by restarting until an attractive run appears.
Proposed outputs
The first deliverables would be small executable examples, technical notes, and an assumption register. Each example would state the graph, initial state, update rule, stopping criterion, and expected failure cases. Comparisons would include communication cost as well as speed.
Any later biological interpretation would require an independently justified mapping from measurable animal behaviour to model variables. Until then, the computational work stands as a theoretical connection and an educational resource.