A finite-dimensional approximation method in optimal control theory (Q1776212)

The authors derive a notion of a (local) Pontryagin minimum for the classical optimal control problem. It is shown that the Pontryagin maximum principle is a necessary condition for a Pontryagin minimum (rather than a usual strong minimum) under very mild smoothness conditions on the functions involved. The proof of the Pontryagin maximum principle is based on the finite-dimensional approximation that was applied for the first time in [\textit{V. M. Alekseev, V.M. Tikhomirov}, and \textit{S. V. Fomin}, ``Optimal control'' (1979; Zbl 0516.49002)] to smooth problems in the class of piecewise continuous controls and later on by the first author in [\textit{A. V. Arutyunov}, ``Extremum conditions. Abnormal and degenerate problems'' (1997; Zbl 0953.49001)] and [\textit{A. V. Arutyunov}, J. Sov. Math. 54, No. 6, 1342--1400 (1991); translation from Itogi Nauki Tekh., Ser. Mat. Anal. 27, 147-235 (1989; Zbl 0726.49015)] to measurable controls in the presence of delays and phase constraints. Other papers in which the Pontryagin maximum principle is analyzed under weaker smoothness conditions are cited.