8 comments

[ 2.9 ms ] story [ 29.3 ms ] thread
I didn't know communication networks could be described with resistors. Are inductors, diodes, and voltage sources also relevant? How far does this go and why would you model networks this way?
[1] has a potential use case. Resistor networks are handy when modeling divergence-free flows through arbitrary networks with linear "costs" from a predefined origin to a predefined destination.

Kirchoff's law and Ohm's law makes things linear, and turning the optimization problem into a linear problem is very advantageous for efficiency.

Diodes are asymmetric resistors with R=0 for one direction and R=inf for the other, so represent one-way streets in the network. A voltage source would be how you define the origin location (destination location is set at ground) for recognizing different routes for flows to take. Inductors are out of scope I think since this should find the equilibrium for voltages at each node and currents across each resistor and inductors' dynamics are only interesting over time.

[1] https://ai.googleblog.com/2022/02/robust-routing-using-elect...

What optimization problem is being solved?
(comment deleted)
Minimal documentation is included here. I'd like to read the associate paper/presentation to get a better sense of its abilities but it hasn't been posted yet.
For all practical purposes, posting that code without an actual guide/readme and without the paper is useless. What are they solving, how is it better than xyz alternative, etc.
Tangentially related: For my masters thesis I developed a framework to model biological reaction networks as electrical circuit elements - concentrations as capacitors, enzymatic activities as (reciprocal of) resistance, etc. This allowed an easy description of the reactions, and then simulations using spice. This was a while back, and good biochem simulation software wasn't available.

My solution was far from perfect, but as a 20-year-old, I was rather proud :D