A Population-Games Application with the Balls-in-Tubes Experiment

摘要

Many multi-agent control strategies have been implemented in simulations, but not as much in physical applications. In this paper, a design and research of multiagent systems control strategies, under several constraints, are presented. The main contribution of this work is the implementation of evolutionary game theory into distributed control systems, throughout different tests in a real application, known as the Balls-in-Tubes experiment. Moreover, results regarding the performance of the closed-loop system are shown, comparing different possible strategies, under a population-games approach. A new methodology for enhancing the control performance is validated by means of various selected key performance indicators. The demonstration of optimal resource distribution with population dynamics is shown with specific designed scenarios, where the characteristics under which the dynamics perform are modified in each test. Furthermore, a comparison among distributed control strategies is presented.