Emergence of specialized Collective Behaviors in Evolving Heterogeneous Swarms

TL;DR

The paper tackles how specialized collective behaviors can emerge in evolving heterogeneous swarms by leveraging phenotypic plasticity to maintain two co-evolving sub-group controllers within a single genome. The approach uses reservoir neural networks for two sub-groups, evolved under a single fitness function that rewards swarm-level gradient-following in an emergent perception task. Validation shows that sub-group interactions produce robust, improved performance, and an online regulatory mechanism further enhances scalability and adaptability across swarm sizes and environments. The results demonstrate that automatic task partition and adaptive regulation can yield more robust swarm behavior without explicit sub-task design, with broad implications for scalable, flexible swarm robotics. The work contributes a task-agnostic framework for emergent specialization and an efficient mechanism for online adaptation that improves performance in variable conditions.

Abstract

Natural groups of animals, such as swarms of social insects, exhibit astonishing degrees of task specialization, useful to address complex tasks and to survive. This is supported by phenotypic plasticity: individuals sharing the same genotype that is expressed differently for different classes of individuals, each specializing in one task. In this work, we evolve a swarm of simulated robots with phenotypic plasticity to study the emergence of specialized collective behavior during an emergent perception task. Phenotypic plasticity is realized in the form of heterogeneity of behavior by dividing the genotype into two components, with one different neural network controller associated to each component. The whole genotype, expressing the behavior of the whole group through the two components, is subject to evolution with a single fitness function. We analyse the obtained behaviors and use the insights provided by these results to design an online regulatory mechanism. Our experiments show three main findings: 1) The sub-groups evolve distinct emergent behaviors. 2) The effectiveness of the whole swarm depends on the interaction between the two sub-groups, leading to a more robust performance than with singular sub-group behavior. 3) The online regulatory mechanism enhances overall performance and scalability.

Figures (7)

• Figure 1: Reservoir Neuron Network controller design.
• Figure 2: The task environment. (a) The scalar map indicating a random instance of our swarm task. (b) Our swarm with different sub-groups (colored red and green) in simulation.
• Figure 3: Experimental setup for optimizing swarm controllers, where an evolutionary algorithm evaluate different genotypes in our swarm simulator (big dashed box). Please note that our genotype encodes two different controllers colored green and white boxes.
• Figure 4: Validation environments. The black striped line indicates the random initialization location of the swarm (similar to \ref{['fig:swarm_env']}). The striped box in (c) indicates the area of random initialization
• Figure 5: Top: The mean$\pm STD$ (line) and average max (dot) fitness over 100 generations (averaged over 10 runs). Bottom: Mean genotype standard deviation (STD) within the population during the best run (per generation). The overall population mean as a black dotted line and each separate reservoir in solid red and green.