Pollination Networks
Distributed SystemsAlgorithmsResearchMachine Learning

Pollination Networks

MSc Thesis — Bartlett, UCL

This thesis project explores pollination networks as a model for digital communication across distributed systems — specifically for Delay Tolerant Networks (DTN) where no end-to-end transmission path exists between nodes.

Concept

When a pollinator comes in contact with a plant, it collects or dissipates pollen. In analogy to data, this is a viable model for digital communication across distributed networks. The proximity-based exchange model enables sensors to remain dormant until a pollination event occurs — reducing hardware and power requirements while enabling finer-grained data collection.

Implementation

London’s bus system was used as the infrastructure, with buses functioning as mobile data carriers. An agent-based simulation tool was developed to examine data flows, and a novel routing algorithm inspired by Adaptive Foraging strategies was created and benchmarked against two alternatives.

The AF algorithm demonstrated superior performance across efficiency, overhead, and average delay metrics.