scientific article; zbMATH DE number 4042652

From MaRDI portal
Publication:3781640





[For part I see the author, ibid., Exp. 358 (1986; Zbl 0625.49001).] Dynamical optimization of a system characterized by an object function (OFC) and a vector differential equation (DE) \(\dot z=f(z,x,t)\) is considered. Variational methods are proposed for the problem. From the Euler-Lagrange DE necessary conditions (NC) for solving extremals are reduced for conditional (bound) optimization, both in the stationary case and for the transient phase. Pontryagin's principle is shown to be NC, if OFC is Hamiltonian. Its connection with Bellman's principle is demonstrated. Linear systems and quadratic OFC take the main part of the paper. Methods of Riccati equations and Ljapunov functions are discussed for them. They are also considered in the presence of Gaussian white noise. Kalman's filter is sought by the transformation to the dual problem and finally to the algebraic Riccati equation.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3781640)