New Algorithms and Results of Numerical Experiments for Solution of Mathematical Programming and Optimal Control Problems

Abstract

From the beginning of the seventies the investigation on constructive theory of optimal control has been carried out under the guidance of the authors in Minsk (Institute of Mathematics of the BSSR Academy of Sciences; Byelorussian State University). Since then a number of new optimization algorithms have been worked out and tested by computer. At first static linear programming problems have been considered, then the suggested methods have been transferred on other mathematical programming problems and dynamic models. Now methods for solution of linear optimal control problems are investigated in detail and the first steps axe made towards constructing methods for solution of nonlinear problems. Below a survey of some results obtained by the members of Minsk Seminar on constructive optimal control methods is given and new algorithms of solving extremal problems are described.KeywordsFeasible SolutionOptimal Control ProblemLinear Programming ProblemAdaptive MethodMain ConstraintThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.