Thermodynamics-inspired modeling of macroscopic swarm states

Abstract

The collective behavior of swarms is extremely difficultto estimate or predict, even when the local agent rules areknown and simple. The presented work seeks to leverage thesimilarities between fluids and swarm systems to generate athermodynamics-inspired characterization of the collective behavior of robotic swarms. While prior works have borrowed toolsfrom fluid dynamics to design swarming behaviors, they haveusually avoided the task of generating a fluids-inspired macroscopic state (or macrostate) description of the swarm. This workwill bridge the gap by seeking to answer the following question:is it possible to generate a small set of thermodynamics-inspiredmacroscopic properties that may later be used to quantify allpossible collective behaviors of swarm systems? In this paper,we present three macroscopic properties analogous to pressure,temperature, and density of a gas, to describe the behavior ofa swarm that is governed by only attractive and repulsive agentinteractions. These properties are made to satisfy an equationsimilar to the ideal gas law, and also generalized to satisfy thevirial equation of state for real gases. Finally, we investigate howswarm specifications such as density and average agent velocityaffect the system macrostate.