Population Protocols with Unreliable Communication

Population protocols are a model of distributed computation intended for the study networks independent computing agents with dynamic communication structure. Each agent has finite number states, and opportunities occur nondeterministically, allowing involved to change their states based on each other's states. In present paper we unreliable models population variations from point view expressive power. We effects message loss. show that general definition constant-storage such can only compute predicates computable by immediate observation (sometimes also called one-way protocols). Immediate inherently tolerant keep power under wide range fairness conditions. prove large class message-based generally more than becomes strictly less in case.