Generalized Particle Systems: A Locality-Factorized Metamodel for Self-Propelled Particles and Agent-Based Models

We propose a single formal object, the _generalized particle system_ (GPS), under which systems of self-propelled particles (SPPS) and agent-based models (ABMs) are instances of one construction. A GPS is a typed finite population, a per-particle state space of arbitrary internal structure, a global environment, a neighbourhood map that may depend on the current configuration, and type-indexed local update rules; these data assemble into a generator whose iteration is a monoid action in the sense of autonomous dynamical systems. Optional extensions—update schedules and stochastic local rules—are given at the same level of generality, so that sequential and noisy models (the original Primordial Particle System, the Vicsek model) are covered rather than approximated. The construction isolates the one ingredient that separates the two communities: in the classical ABM reading of [1] the dependency graph is fixed, whereas in an SPPS it is a function of the state. We then introduce an _exchangeability_ axiom—rules depend on neighbours only through the multiset of their types and states—and prove that under it the generator is equivariant for type-preserving permutations and descends to a dynamical system on finite counting measures, which is the form in which kinetic and mean-field limits of swarming models are posed. We argue that the tuple should be read as a _metamodel_: its structural components (state space, neighbourhood map, rule family) coordinatize a space of models, inside which the usual numerical parameters live. Six pre-registered numerical experiments, run on a single engine that is a literal implementation of the generator, test the claims: freezing the neighbourhood map destroys collective alignment (\(\varphi = 0.91\rightarrow 0.04\)) while producing a different, non-aligned aggregation; the realized neighbourhood graph rewires at up to \(64\%\) of edges per step and its rewiring rate is anti-correlated with order (\(\text{corr} = {- 0.99}\)); exchanging a metric for a topological neighbourhood measurably reshapes the phase diagram of a flocking rule; the same rule with cohesion and separation switched off exhibits a Vicsek-type order–disorder transition whose susceptibility peak grows with system size; and an internal activation variable does measurable but modest causal work. One pre-registered prediction failed and one subsumption claim could not be confirmed; both are reported.

Authors

Institutions

Publication Details

Journal
Qeios
Published
2026-11-17
DOI
https://doi.org/10.32388/ho6iov
Primary Topic
Distributed Control Multi-Agent Systems
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Generalized Particle Systems: A Locality-Factorized Metamodel for Self-Propelled Particles and Agent-Based Models

Igor Strozzi
Qeios
Distributed Control Multi-Agent Systems
preprint

Generalized Particle Systems: A Locality-Factorized Metamodel for Self-Propelled Particles and Agent-Based Models

Igor Strozzi
preprint en

Abstract

We propose a single formal object, the _generalized particle system_ (GPS), under which systems of self-propelled particles (SPPS) and agent-based models (ABMs) are instances of one construction. A GPS is a typed finite population, a per-particle state space of arbitrary internal structure, a global environment, a neighbourhood map that may depend on the current configuration, and type-indexed local update rules; these data assemble into a generator whose iteration is a monoid action in the sense of autonomous dynamical systems. Optional extensions—update schedules and stochastic local rules—are given at the same level of generality, so that sequential and noisy models (the original Primordial Particle System, the Vicsek model) are covered rather than approximated. The construction isolates the one ingredient that separates the two communities: in the classical ABM reading of [1] the dependency graph is fixed, whereas in an SPPS it is a function of the state. We then introduce an _exchangeability_ axiom—rules depend on neighbours only through the multiset of their types and states—and prove that under it the generator is equivariant for type-preserving permutations and descends to a dynamical system on finite counting measures, which is the form in which kinetic and mean-field limits of swarming models are posed. We argue that the tuple should be read as a _metamodel_: its structural components (state space, neighbourhood map, rule family) coordinatize a space of models, inside which the usual numerical parameters live. Six pre-registered numerical experiments, run on a single engine that is a literal implementation of the generator, test the claims: freezing the neighbourhood map destroys collective alignment (\(\varphi = 0.91\rightarrow 0.04\)) while producing a different, non-aligned aggregation; the realized neighbourhood graph rewires at up to \(64\%\) of edges per step and its rewiring rate is anti-correlated with order (\(\text{corr} = {- 0.99}\)); exchanging a metric for a topological neighbourhood measurably reshapes the phase diagram of a flocking rule; the same rule with cohesion and separation switched off exhibits a Vicsek-type order–disorder transition whose susceptibility peak grows with system size; and an internal activation variable does measurable but modest causal work. One pre-registered prediction failed and one subsumption claim could not be confirmed; both are reported.

Qeios
Universidade Federal do Rio de Janeiro (BR)
Distributed Control Multi-Agent Systems
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon