arXiv subjects

Mihir Durve

Publications and source records attributed to Mihir Durve.

22 records · Page 2Linked to original sources

arXiv2019

Collective olfactory search in a turbulent environment

Finding the distant source of an odor dispersed by a turbulent flow is a vital task for many organisms, either for foraging or for mating purposes. At the level of individual search, animals like moths have developed effective strategies to solve this very difficult navigation problem based on the noisy detection of odor concentration and wind velocity alone. When many individuals concurrently perform the same olfactory search task, without any centralized control, sharing information about the decisions made by the members of the group can potentially increase the performance. But how much of this information is actually valuable and exploitable for the collective task ? Here we show that, in a model of a swarm of agents inspired by moth behavior, there is an optimal way to blend the private information about odor and wind detections with the publicly available information about other agents' heading direction. At optimality, the time required for the first agent to reach the source is essentially the shortest flight time from the departure point to the target. Conversely, agents who discard public information are several fold slower and groups that do not put enough weight on private information perform even worse. Our results then suggest an efficient multi-agent olfactory search algorithm that could prove useful in robotics, for instance in the identification of sources of harmful volatile compounds.

physics.bio-ph↗

arXiv2019

Learning to flock through reinforcement

Flocks of birds, schools of fish, insects swarms are examples of coordinated motion of a group that arises spontaneously from the action of many individuals. Here, we study flocking behavior from the viewpoint of multi-agent reinforcement learning. In this setting, a learning agent tries to keep contact with the group using as sensory input the velocity of its neighbors. This goal is pursued by each learning individual by exerting a limited control on its own direction of motion. By means of standard reinforcement learning algorithms we show that: i) a learning agent exposed to a group of teachers, i.e. hard-wired flocking agents, learns to follow them, and ii) that in the absence of teachers, a group of independently learning agents evolves towards a state where each agent knows how to flock. In both scenarios, i) and ii), the emergent policy (or navigation strategy) corresponds to the polar velocity alignment mechanism of the well-known Vicsek model. These results show that a) such a velocity alignment may have naturally evolved as an adaptive behavior that aims at minimizing the rate of neighbor loss, and b) prove that this alignment does not only favor (local) polar order, but it corresponds to best policy/strategy to keep group cohesion when the sensory input is limited to the velocity of neighboring agents. In short, to stay together, steer together.

physics.soc-ph↗

arXiv2018

Active Particle Condensation by Nonreciprocal and Time-delayed Interactions

We consider flocking of self-propelling agents in two dimensions, each of which communicates with its neighbors within a limited vision cone. Also, the communication occurs with some delay. The communication among the agents are modeled by Vicsek rules. In this study we explore the effect of non-reciprocal interaction among the agents, induced by their vision cone, together with the delayed interactions on the dynamical pattern formation within the flock. We find that under these two influences and without any position based attractive interactions or confining boundaries, the agents can spontaneously condense into drops. Though the agents are in motion within the drop, the drop as whole is virtually pinned in space. We also find that this novel state of the flock has a well defined order stabilized by the noise present in the system.

cond-mat.soft↗

arXiv2016

First order phase transition in a self-propelled particles model with variable angular range of interaction

We have carried out a Monte Carlo simulation of a modified version of Vicsek model for the motion of self-propelled particles in two dimensions. In this model the neighborhood of interaction of a particle is a sector of the circle with the particle at the center (rather than the whole circle as in the original Vicsek model). The sector is centered along the direction of the velocity of the particle, and the half-opening angle of this sector is called the `view-angle'. We vary the view-angle over its entire range and study the change in the nature of the collective motion of the particles. We find that ordered collective motion persists down to remarkably small view-angles. And at a certain critical view-angle the collective motion of the system undergoes a first order phase transition to a disordered state. We also find that the reduction in the view-angle can in fact increase the order in the system significantly. We show that the directionality of the interaction, and not only the radial range of the interaction, plays an important role in the determination of the nature of the above phase transition.

cond-mat.stat-mech↗