Taylor series method for dynamical systems with control: convergence and error estimates (Q2462050)

The authors consider an initial-value problem for a first order system of differential equations having polynomial right-hand sides with respect to all dependent variables. First, the Taylor series method for this system is discussed: the true solution of the system is approximated by the partial sum of the Taylor series of a given order and then, formulas for calculating the Taylor coefficients are given. The step size and the guaranted local error estimates are obtained. In the next section, explicit Runge-Kutta type methods are applied to the differential system and the step size and error estimates are also derived in this case. For a more precise solution, the authors introduce the so-called ``aggregative Taylor series'' method as a result of many steps of the same size of the Taylor series method. The conclusions of the authors are that the Taylor series method and the aggregative Taylor series method are both faster and more precise than the explicit Runge-Kutta methods of comparable orders. Practical implementations of the methods are proposed and many type of numerical examples are performed.