We report about the application of Reinforcement Learning to optimize tasks of single agents (swimmers/vessels) in 2d and 3d complex flows, by studying a series of paradigmatic cases: objects that can regulate their buoyancy [1], micro-swimmers with preferential adjustable velocity direction [2] and small drifter able to regulate their slip velocity with respect to the advecting 2d turbulent flow to perform point-to-point navigation in optimal time [3].
[1] S. Colabrese, K. Gustavsson, A. Celani and L. Biferale. Smart Inertial Particles. Phys. Review Fluids 3, 084301 (2018)
[2] S. Colabrese, K. Gustavsson, A. Celani and L. Biferale. Flow navigation by smart microswimmers via reinforcement learning. Phys. Rev. Lett. 118 (15), 158004 (2017)
[3] L. Biferale, F. Bonaccorso, M. Buzzicotti, P. Clark Di Leoni and K. Gustavsson. Point-to-point optimal navigation for the Zermelo problem in 2d turbulent flows using Reinforcement Learning. (in preparation)