Optimal Control Problem of Switched Systems (OCPSS) consists in determining the control variable profile that minimizes a given objective function subject to differential-algebraic constraints defined by phases. Traditionally, this problem has been solved using classical techniques through algebraic manipulation of original problem (obtaining the sensitivity equations) and resolution, at each step, a boundary value problem highly dependent of initial estimates. In this context, the present work aim to resolve the OCPSS using the Gravitational Search Algorithm. This search strategy is based on the Newtonian law of gravity for the generation of potential candidates to solve optimization problems. The results are compared with those obtained by the Levenberg-Marquardt Algorithm and with a hybrid version.
Optimal control; switched differential Algebraic-Equation; Levenberg-Marquardt algorithm; gravitational search algorithm; hybridization
