Stabilization of Nonholonomic System Using Evolutionary Strategies

The minimization of the quadratic integral approach, in the optimal control, is a measure that determines that so quick an initial state is taken to the state zero, considering that its parameters are: the coordinates of the dynamic system, the information of the control and positive defined matrices. These positive defined matrices are obtained in particular based on the engineer´s experience on a dynamic system. It is clear that, the good election of the matrices, in an enclosed group, they should improve the objective of the approach. It is for it that, we propose an algorithm, based on evolutionary strategies, to choose the positive defined matrices, in an enclosed space. The important result is given in the operations of crossover and mutation of the evolutionary algorithm: they preserve the positive defined properties of the matrices, improving the computational complexity of this algorithm.