Stable information transfer network facilitates the emergence of collective behavior of bird flocks

Abstract

Collective behavior is ubiquitous in living systems such as human crowds, microbial communities, insect swarms, and bird flocks. It is widely reported that simple, local, temporal interactions between individuals may lead to the emergence of collective behavior. In large-scale flocks, however, the global information transfer in dynamical models with time-varying neighborhoods is not normally instantaneous. Here, we present an information transfer network in which interactions last for a certain period of time. We then combine this network with a dynamic model of self-propelled particles. We observe that the resulting model that uses a stable information transfer network creates a bird flock with a more robust performance. We further show that the time of information transfer in a flock grows logarithmically with its size and is proportional to the average response time of the birds. Moreover, we find that the ranking curves displaying the order in which birds first perceive an external stimulus have similar shapes across different flocks. Our results demonstrate that, beyond the traditional local, temporal interactions, our stable information transfer network serves as an efficient mechanism to model the emergence of large-scale collective behavior.

Collective turns in jackdaw flocks: kinematics and information transfer

“…Consequently, many studies have focused on inferring the interaction rules followed by birds [31][32][33][34][35][36][37][38][39][40] and their implications for macroscopic quantities such as flock morphology [32,38,41] and the spatial distance over which the velocity fluctuations of individuals are correlated [37,42,43]. Flocking behaviour, including collective turns, has also been numerically simulated using discrete selfpropelled particles [44,45], network models [46] and continuum hydrodynamic models [47,48]. Although there have been a number of empirical studies on the collective turns of large groups of birds in the wild [14,18,21,49,50], many questions remain outstanding.…”

“…turns, has also been numerically simulated using discrete selfpropelled particles [44,45], network models [46] and continuum hydrodynamic models [47,48]. Although there have been a number of empirical studies on the collective turns of large groups of birds in the wild [14,18,21,49,50], many questions remain outstanding.…”

Emergence of splits and collective turns in pigeon flocks under predation

“…If fewer than half of the flock follows the initiator, a sub-flock splits off (which constitutes approximately 10% of all splits in our simulated data). According to rules of information propagation through a group [ 36 , 66 , 81 ], we hypothesized that more collective turns will be initiated by individuals that are turning slowly (with low angular velocity) and inwards, and are positioned close to the centre of their flock.…”

“…We show that both splitting and collective turning emerge in our model from a single escape manoeuvre. Based on theories of turning propagation [ 36 , 64 , 65 ] and information transfer [ 36 , 66 ], we expect collective turns to emerge more frequently when initiators: (i) are positioned more centrally in the group (directional information can reach the edges of the group faster), (ii) are turning slower and (iii) turn towards the group's centroid (making it easier for flock-mates to follow). We test these hypotheses in both our model and empirical data .…”

Controlling Network Dynamics

“…There are indeed experiments conducted over each pairwise community to infer the underlying networks of several species [93], but challenges still remain for large systems as well as higher-order interactions for both microbial communities [94,95] and social systems [96,97]. On the whole, for self-organized systems like the flocking behavior of a huge number of selfpropelled birds without central coordination [98][99][100], or the massive and diverse microbes dwelling on almost every surface of our body [54,[101][102][103], more efficient algorithms and methods are aspired to extrapolate the underlying skeleton precisely.…”