Poiseuille Flow Assumption: The Hydrodynamic Roots of Slime Algorithms

Physarum-inspired shortest-path algorithms treat each network edge like a tube carrying flow. The core assumption is Poiseuille-like behavior: flux rises with pressure difference and conductivity, and falls with edge length.

A standard form is:

Q_ij = (D_ij / L_ij) * (p_i - p_j)

This one equation is the backbone of the algorithmic adaptation loop.

What each variable means

• Q_ij: flux on edge i-j
• D_ij: edge conductivity, adaptive capacity to carry flow
• L_ij: edge length or traversal cost surrogate
• p_i - p_j: pressure difference across nodes

If pressure drop is strong and conductivity is high, flux rises. If length is large, flux is damped.

Why this assumption matters algorithmically

The model creates a natural positive feedback cycle.

1. Solve pressures on the current graph.
2. Compute flux on edges.
3. Increase conductivity on high-flux edges.
4. Decrease conductivity on low-flux edges.

After repeated iterations, irrelevant routes fade and efficient routes remain.

What this captures well

Poiseuille-style flux modeling captures continuous adaptation, not one-shot discrete path selection. That is why Physarum algorithms can preserve multiple competitive routes before converging.

It also links geometry and transport naturally: longer edges become expensive unless pressure support is strong.

What it simplifies

The model is a mathematical abstraction, not a full biological replica. Real plasmodial flow includes nonlinear biochemical regulation and active deformation dynamics.

Still, as a computational principle for decentralized routing, the assumption is strong and productive.

Related reading: Kirchhoff’s Biology, Poisson Pressure Solver, and IPPA Energy Functions.

Origin and E-E-A-T

This article is based on Source 16 notes in the project source archives for Zhang et al.’s improved Physarum algorithm, where flux is modeled with a Poiseuille-style relation between pressure, conductivity, and edge length. Editorial review completed on 2026-02-11, version 1.0.0.