Hunting active Brownian particles: Learning optimal behavior

Abstract

We numerically study active Brownian particles that can respond to environmental cues through a small set of actions (switching their motility and turning left or right with respect to some direction) which are motivated by recent experiments with colloidal self-propelled Janus particles. We employ reinforcement learning to find optimal mappings between the state of particles and these actions. Specifically, we first consider a predator-prey situation in which prey particles try to avoid a predator. Using as reward the squared distance from the predator, we discuss the merits of three state-action sets and show that turning away from the predator is the most successful strategy. We then remove the predator and employ as collective reward the local concentration of signaling molecules exuded by all particles and show that aligning with the concentration gradient leads to chemotactic collapse into a single cluster. Our results illustrate a promising route to obtain local interaction rules and design collective states in active matter.

arXiv commentto appear in Phys. Rev. E

References in corpus (30)

• Active Particles in Complex and Crowded Environments
• Collective motion
• Self-motile colloidal particles: from directed propulsion to random walk
• Physics of Microswimmers - Single Particle Motion and Collective Behavior
• Intermittent search strategies
• Dynamic clustering in active colloidal suspensions with chemical signaling
• The 2019 Motile Active Matter Roadmap
• Optimal search strategies for hidden targets
• Phototaxis of synthetic microswimmers in optical landscapes
• Gravitaxis of asymmetric self-propelled colloidal particles
• Flow Navigation by Smart Microswimmers via Reinforcement Learning
• Dynamic clustering and chemotactic collapse of self-phoretic active particles
• When does a physical system compute?
• Learning to school in the presence of hydrodynamic interactions
• Active Particles Bound by Information Flows
• Viscotaxis: microswimmer navigation in viscosity gradients
• Active particles in viscosity gradients
• Optimal steering of a smart active particle
• Chemotactic predator-prey dynamics
• Self-learning how to swim at low Reynolds number
• Microswimmers learning chemotaxis with genetic algorithms
• Learning to flock through reinforcement
• Active Brownian motion with orientation-dependent motility: theory and experiments
• Learning to Control Active Matter
• Active-Particle Polarization Without Alignment Forces
• Quorum-sensing active particles with discontinuous motility
• Markovian robots: minimal navigation strategies for active particles
• A single predator charging a herd of prey: effects of self volume and predator-prey decision-making
• Chemokinetic scattering, trapping, and avoidance of active Brownian particles
• Numerical analysis of homogeneous and inhomogeneous intermittent search strategies

Cited by in corpus (3)

• Reinforcement learning of optimal active particle navigation
• Noisy Pursuit and Pattern Formation of Self-Steering Active Particles
• Autotaxis of a Run-and-Tumble Agent via Reinforcement Learning