Computational algorithm for solving problem of optimal boundary-control with nonseparated boundary conditions

Abstract

The problem of optimal boundary-control with nonseparated boundary conditions is considered in the case where the motion of a plant is described by a system of nonlinear ordinary differential equations with a corresponding initial condition, which thereafter is taken as a control action. First, the system of corresponding nonlinear Euler---Lagrange equations is described and in order to solve it, the quasi-linearization method (the first method) is taken. After that, the quasi-linearization method is used with the aim to solve just the optimization problem with boundary control and the nonseparated boundary condition; as a result, the initial nonlinear problem is reduced to the solution of a corresponding linearly quadratic optimization problem (the second method). By the particular example (of oil production) we demonstrate that the second method converges considerably faster and its accuracy is five times higher than the accuracy of the first method. The numerical results reinforce the compliance of the constructed mathematical model with practice.