State-control spectral Chebyshev parameterization for linearly constrained quadratic optimal control problems (Q678808): Difference between revisions

For linearly constrained quadratic optimal control problems of the form: Maximize or Minimize \[ \begin{multlined} J=\textstyle {1\over 2} x^T(t_f)H \bigl(x(t_f)+h^Tx(t_f)+\textstyle {1\over 2}\int^{t_f}_0 x^T(t)Q(t)x(t)+u^T(t)R(t)u(t)+\\ +x^T(t)S(t)u(t)+q^T(t)x(t)+r^T(t) u(t)\bigr)dt\end{multlined} \] subject to \(\dot x(t)= A(t)x(t)+B(t)u(t)\), \(t\in[0,t_f]\), and \(x(0)=x_0\), \(E_1(t) x(t)+ E_2(t) u(t)\leq e(t)\), the author gives a direct computational method, which is based on a cell averaging method in which the \(m\)th degree interpolating polynomial is constructed, using Chebyshev nodes. The given problem is thereby transformed into a quadratic programming one. Illustrative examples are given.