A Reservoir Computer-Based Modeling of Hunting Dynamics in Predator-Prey Scenarios

Several optimization algorithms such as particle swarm or evolutionary algorithms are inspired by biological processes. For instance, the hunting dynamics of predator-prey scenarios have been identified to model dynamical optimization problems. Inspired by this, we develop an ideal electrical circuit in combination with a delay-coupled reservoir computer as a model for each animal in the scenario. This reservoir computer accounts for the visual processing of the animal, while an RLC circuit and an RL circuit model the calculation of the animal’s velocity and its position, respectively. Here, the relative position of predator and prey is fed back into the reservoir computer. Simulation results for two environments with different terrain slopes show that this setup indeed mimics a predator chasing preys, where the choice of the hunted prey is strongly affected by the terrain.