Informatica

We study the stochastic model for bioremediation in a bioreactor with ideal mixing. The dynamics of the examined system is described by stochastic differential equations. We consider an optimal control problem with quadratic costs functional for the linearized model of a well-stirred bioreactor. The optimal control is based on the optimal robust state estimates. The approximate optimal solution is obtained as a linear feedback.

This paper deals with the study of the optimal control problem for the objective F(x,u,v)=∫^T0f(x(t),x(t−h),ζ(t),u(t),v(t),t)dt, with x∈X,u∈U,v∈V;X,U and V being vector spaces, and ζ(t)=∫^h0R(t,τ)x(t−τ)dτ subject to the differential equation $\frac{d}{dt}x(t)=m(x(t),x(t-h),\zeta(t),u(t),v(t),t)(0\leq t\leq T)$, and the constraints g1(u(t),t)∈S1,g2(v(t),t)∈S2;n1(x(t),t)∈V1,n2(x(t−h),t)∈V2;n3(ζ(t),t)∈V3(0≤t≤T), where x(t)∈Rⁿ;ζ(t)∈Rⁿ;u(t)∈R^k;f,m,gi(i=1,2),ni(1≤i≤3) and the entries to r(t,τ):R+×R+→L(X,X) are continuously differentiable functions. It is assumed that boundary conditions x(0)=x(T)=0 are imposed. Si(i=1;2) and Vi(1≤i≤3) are convex cones. The existence of a time-optimal control in analytic linear systems is also investigated via an extension of the bang-bang principle.

We consider a possibility of automating the analysis of a computer program realizing the objective function of an extremal problem, and of distributing the calculation of the function value into parallel processes on the basis of results of the analysis. The first problem is to recognize the constituent parts of the function. The next one is to determine their computing times. The third problem is to distribute the calculation of these parts among independent processes. A special language similar to PASCAL has been used to describe the objective function. A new scheduling algorithm, seeking to minimize the maximal finishing time of processing units, was proposed and investigated. Experiments are performed using a computer network.

Error bounds are developed for a class of quadratic programming problems. The absolute error between an approximate feasible solution, generated via a dual formulation, and the true optimal solution is measured. Furthermore, these error bounds involve considerably less work computationally than existing estimates.

A new concept of an exact auxiliary function (EAF) is introduced. A function is said to be EAF, if the set of global minimizers of this function coincides with the global solution set of the initial optimization problem. Sufficient conditions for exact equivalence of the constrained minimization problem and minimization of EAF are provided. The paper presents various classes of EAF for a non linear programming problem, which has a saddle point of Lagra ge function.