On a minimax control problem for a positional functional under geometric and integral constraints on control actions

Within the game-theoretical approach, we consider a minimax feedback control problem for a linear dynamical system with a positional quality index, which is the norm of the deviation of the motion from given target points at given times. Control actions are subject to both geometric and integral constraints. A procedure for the approximate calculation of the optimal guaranteed result and for the construction of a control law that ensures the result is developed. The procedure is based on the recursive construction of upper convex hulls of auxiliary program functions. Results of numerical simulations are presented.     

Numerical solution of a minimax ergodic optimal control problem

In this work we consider an L^∞ minimax ergodic optimal control problem with cumulative cost. We approximate the cost function as a limit of evolutions problems. We present the associated Hamilton-Jacobi-Bellman equation and we prove that it has a unique solution in the viscosity sense. As this HJB equation is consistent with a numerical procedure, we use this discretization to obtain a procedure for the primitive problem. For the numerical solution of the ergodic version we need a perturbation of the instantaneous cost function. We give an appropriate selection of the discretization and penalization parameters to obtain discrete solutions that converge to the optimal cost. We present numerical results. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

On the numerical solution of Plateau's problem

The present paper is dedicated to the numerical computation of minimal surfaces by the boundary element method. Having a parametrization γ of the boundary curve over the unit circle at hand, the problem is reduced to seeking a reparametrization κ of the unit circle. The Dirichlet energy of the harmonic extension of γ○κ has to be minimized among all reparametrizations. The energy functional is calculated as boundary integral that involves the Dirichlet-to-Neumann map. First and second order necessary optimality conditions of the underlying minimization problem are formulated. Existence and convergence of approximate solutions is proven. An efficient algorithm is proposed for the computation of minimal surfaces and numerical results are presented.     

On the minimax control problem for nonlinear operator-differential equations

We consider the problem of minimax control for objects describable by nonlinear operator-differential equations in Banach spaces with restrictions on the phase variables and controls. We prove solvability theorems. The method of proof is based on the theory of extremal problems for generalized operator-differential equations.Translated fromVychislitel'naya i Prikladnaya Matematika, Issue 71, 1990, pp. 114–119.     

A numerical method for the minimax solution of the Bellman equation in the Cauchy problem with additional restrictions

A new numerical method is suggested for constructing the value function in optimal control problems of prescribed duration with positional running cost along motions of controlled dynamical systems. The algorithm is based on a backward procedure involving characteristics of the Bellman equation. Estimations of the approximation are provided. Results of simulations for a model example are exposed.     

Infinite-horizon minimax control with pointwise cost functional

This paper is devoted to the discussion of a minimax optimal control problem over an infinite-time horizon, where the functional to be minimized is the highest instantaneous cost that may occur under the worst combination of disturbances. The problem is formulated for a general stationary discrete-time dynamic system, and a dynamic programming algorithm is proposed for its solution. The relationship existing between the cost functional associated to a control law and the reachability properties of the resulting controlled system is discussed.     

Infinite-horizon periodic minimax control problem

This note analyzes a minimax optimal control problem involving a periodic discrete-time system. The functional to be minimized is the maximum value of the sum over a period of the costs incurred by the system. A dynamic programming algorithm is proposed for the solution.     

On the numerical solution of the linear complementarity problem

The well-known linear complementarity problem with definite matrices is considered. It is proposed to solve it using a global optimization algorithm in which one of the basic stages is a special local search. The proposed global search algorithm is tested using a variety of randomly generated problems; a detailed analysis of the computational experiment is given.     

A minimax optimal control problem

A control system x=f(t,x,u) is considered, and a cost functional ess supT ₀tT ₁ G(t, x(t),u(t)) is to be minimized. Necessary conditions for optimality (maximum principle and transversality conditions) are derived. It is also shown that an optimal control is optimal for the corresponding problem on a subinterval of [T ₀,T ₁], if a certain controllability condition is satisfied.     

On the numerical solution of the quadratic eigenvalue complementarity problem

The Quadratic Eigenvalue Complementarity Problem (QEiCP) is an extension of the Eigenvalue Complementarity Problem (EiCP) that has been introduced recently. Similar to the EiCP, the QEiCP always has a solution under reasonable hypotheses on the matrices included in its definition. This has been established in a previous paper by reducing a QEiCP of dimension n to a special 2n-order EiCP. In this paper we propose an enumerative algorithm for solving the QEiCP by exploiting this equivalence with an EiCP. The algorithm seeks a global minimum of a special Nonlinear Programming Problem (NLP) with a known global optimal value. The algorithm is shown to perform very well in practice but in some cases terminates with only an approximate optimal solution to NLP. Hence, we propose a hybrid method that combines the enumerative method with a fast and local semi-smooth method to overcome the latter drawback. This algorithm is also shown to be useful for computing a positive eigenvalue for an EiCP under similar assumptions. Computational experience is reported to demonstrate the efficacy and efficiency of the hybrid enumerative method for solving the QEiCP.     

Functional and numerical solution of a control problem originating from heat transfer

The solution of a nonlinear parabolic equation is studied from both the functional and the numerical points of view. Existence, uniqueness, and stability results are given. A boundary-control problem is then presented. Expressions of the gradient and the Hessian of the cost function are given with some details, and the Newton method is compared with a gradient method and the Fletcher-Reeves method. Numerical results are given. The paper concludes with a discussion of the advantages of the Newton method as applied to a control problem.     

On an optimal control problem with an integral functional of a rational control function

We present the solution of an optimization problem with integral performance functional that is nonlinear with respect to the control and contains a discounting parameter in the class of programmed controls under two-sided control constraints. The optimal control is found in the form of a function of time (a program). On the basis of the theoretical results, we perform numerical experiments with model and real data.     

On a class of constrained functional minimization problems and their numerical solution

We investigate a class of functional minimization problems with constraints. By means of variational principles, optimal control theory, and numerical methods for nonlinear equations, numerical methods and the corresponding computer software are established to solve the problems. These tools can be used in fitting curves with arbitrary smoothness, different boundary conditions, and constraints. For special boundary conditions, analytical expressions of the curves are derived. Numerical examples are given to demonstrate the effectiveness of the algorithms by the means of curve fitting.     

On the numerical solution of a nonlocal boundary value problem for a degenerating pseudoparabolic equation

We study a nonlocal boundary value problem for a degenerating pseudoparabolic third-order equation of the general form. For the solution of the problem, we obtain a priori estimates in differential and difference form, which imply the stability of the solution with respect to the initial data and right-hand side on a layer as well as the convergence of the solution of the difference problem to the solution of the differential problem.     

Solution approximation in a minimax control problem for a singularly perturbed system with delay

We consider a control problemwith a minimax criterion for a singularly perturbed system with delay. We propose a procedure for constructing an initial approximation to control in the minimax control problem.     

The problem of computing the value of a differential game for a positional functional

A dynamical system driven by controls and uncontrollable noise is considered in a game-theoretic setting [1–8]. The problem of feedback control in which the performance index is a positional functional of the motion of the system [8–11] is investigated. On the assumption that the structure of the functional satisfies reasonably general conditions, a procedure is proposed for computing the value of the corresponding differential game. Irrespective of the number of dimensions in the initial problem, as dictated by the structure of the performance index, the proposed procedure reduces to the problem of the successive construction of the upper convex hulls of certain auxiliary functions in domains whose dimension does not exceed that of the phase vector of the system.