Enhancing the performance of a bistable energy harvesting device via the cross-entropy method

This work deals with the solution of a non-convex optimization problem to enhance the performance of an energy harvesting device, which involves a nonlinear objective function and a discontinuous constraint. This optimization problem, which seeks to find a suitable configuration of parameters that maximize the electrical power recovered by a bistable energy harvesting system, is formulated in terms of the dynamical system response and a binary classifier obtained from 0 to 1 test for chaos. A stochastic solution strategy that combines penalization and the cross-entropy method is proposed and numerically tested. Computational experiments are conducted to address the performance of the proposed optimization approach by comparison with a reference solution, obtained via an exhaustive search in a refined numerical mesh. The obtained results illustrate the effectiveness and robustness of the cross-entropy optimization strategy (even in the presence of noise or in moderately higher dimensions), showing that the proposed framework may be a very useful and powerful tool to solve optimization problems involving nonlinear energy harvesting dynamical systems.