ABSTRACT

The problems of optimal control with non-separated multipoint and integral conditions are investigated. It is proposed to use the optimization methods of the first order, applying the formulas of the functional gradient derived in the work, for the numerical solution. To solve the direct and conjugate boundary value problems, an approach that is based on convolution operation of integrated conditions into the non-separated local ones and their consecutive shift, which is analogous to the conditions transfer, is proposed. This approach allows one to reduce the solution of the initial boundary value problems to solving the auxiliary Cauchy problem and system of linear algebraic equations. The results of the experiments are given.