Approximation by Micchelli Combination of Half Bernstein Polynomials

A positive operator on a cone is a linear operator that maps the cone to a subcone of itself. The extended second order cones were introduced by Nemeth and Zhang [17] as a working tool to solve mixed complementarity problems. Sznajder [23] determined the automorphism group and the Lyapunov (or bilinearity) ranks of these cones. Ferreira and Nemeth [9] reduced the problem of projecting onto the second order cone to a piecewise linear equation. Nemeth and Xiao [16] solved linear complementarity problems on the extended second order cone (motivated by portfolio optimization models) by reducing them to mixed complementarity problems with respect to the nonnegative orthant. As an extension of Sznajder's results, this paper aims to be a first work about finding necessary conditions and sufficient conditions for a linear operator to be a positive operator (which extends the notion of an automorphism) on an extended second order cone. Although, in the particular case of second order cones a necessary and sufficient condition is known, for extended second order cone such a condition is very difficult to find without restricting the structure of the linear operator. If the matrix of the linear operator is block-diagonal, we give such a necessary and sufficient condition.

For a sequence of positive linear approximation operators defined by means of the squared Bernstein polynomials we derive a complete asymptotic expansion. The initial three coefficients are explicitly calculated. As a special case, we obtain a Voronovskaja type formula. Finally, we introduce two Durrmeyer-type variants and calculate the initial two coefficients of their asymptotic expansions.

We provide an analysis of the rate of convergence of positive linear approximation operators defined on C[0, 1]. We obtain a sufficient condition for a sequence of positive linear approximation operators to possess a Mamedov-type property and give an application to the Durrmeyer approximation process.

It is proved that at almost all points the order of approximation, even of one of the functions 1, cos x,sin x by means of a sequence of linear positive polynomial operators having uniformly bounded norms, is not higher than 1/n2. Refinements of this result are given for operators of convolution type.

Recently there has been a great amount of research both qualitative and quantitative on the problem of approximating functions with certain growth conditions by means of linear positive operators. This paper discusses a quantitative problem for unbounded functions, i.e., the order of approximation to unbounded functions by means of linear positive operators. The theorems presented are, in several instances, extensions of known results from the case of bounded functions to the unbounded case. Applications of these theorems are given to spline functions, which are presently the subject of intensive investigation, and to the well-known Szasz-Mirakyan-Hille operators.

In this work we consider quantitatively with rates the convergence of sequences of linear operators applied on Banach space valued functions. The results are pointwise estimates with rates. To prove our main results we use an elegant and natural boundedness property of our linear operators by their companion positive linear operators. Our inequalities are generalized g-direct and iterated fractional involving the right and left vector Caputo type generalized g-direct and iterated fractional derivatives, built in vector moduli of continuity. We treat wide and general classes of Banach space valued functions. We give applications to vectorial Bernstein operators. See also [6].

Throughout this paper we denote by L p the Banach lattice of p-integrable functions on a σ-finite measure space (X, B, μ), where 1 ≤ p ≤ ∞. We will consider those aspects of the theory of positive linear operators, which are in some way special due to the fact the operators are acting on L p-spaces. For general information about positive operators on Banach lattices we refer to the texts [1]. [20], and [36]. Our focus on L p-spaces does not mean that in special cases some of the results can not be extended to a larger class of Banach lattices of measurable function such as Orlicz spaces or re-arrangement invariant Banach function spaces. However in many cases the results in these extensions are not as precise or as complete as in the case of L p-spaces. We will discuss results related to the boundedness of positive linear operators on L p-spaces. The most important result is the so-called Schur criterion for boundedness. This criterion is the most frequently used tool to show that a concrete positive linear operator is bounded from L p to L q. Then we will show how this result relates to the change of density result of Weis [33]. Next the equality case of Schur’s criterion is shown to be closely related to the question whether a given positive linear operator attains its norm. We discuss in detail the properties of norm attaining operators on L p-spaces and discuss as an example the weighted composition operators on L p-spaces. Then we return to the Schur criterion and show how it can be applied to the factorization theorems of Maurey and Nikisin. Most results mentioned in this paper have appeared before in print, but sometimes only implicitly and scattered over several papers. Also a number of the proofs presented here are new.

The paper centers around a pair of sequences of linear positive operators. The former has the degree of exactness one and the latter has its degree of exactness null, but, as a novelty, it reproduces the third test function of Korovkin theorem. Quantitative estimates of the rate of convergence are given in different function spaces traveling from classical approximation to approximation in abstract spaces. Particular classes are also studied.

We provide a data-driven framework for optimal control of a continuous-time control-affine stochastic dynamical system. The proposed framework relies on the linear operator theory involving Perron-Frobenius (P-F) and Koopman operators. Our first result involving the P-F operator provides a convex formulation to the optimal control problem in the dual space of densities. This convex formulation of the stochastic optimal control problem leads to an infinite-dimensional convex program. The finite-dimensional approximation of the convex program is obtained using a data-driven approximation of the linear operators. We provide rigorous error bounds and convergence rate for the data-driven approximation of the linear operators and comment on the convergence rate of the optimal control w.r.t. data size. Our second results demonstrate using the Koopman operator, dual to the P-F operator, for the stochastic optimal control design. We show that the Hamilton Jacobi Bellman (HJB) equation can be expressed using the Koopman operator. We provide an iterative procedure along the lines of a popular policy iteration algorithm based on the data-driven approximation of the Koopman operator for solving the HJB equation. The two formulations, namely the convex formulation involving the P-F operator and the Koopman-based formulation using the HJB equation, can be viewed as dual to each other where the duality follows due to the dual nature of the P-F and Koopman operators. Finally, we present examples to demonstrate the efficacy of the developed framework and verify the convergence rates for the operator and optimal control numerically.

Let Ω be a compact convex domain in \({\mathbb{R}}^{d}\) and let L be a bounded linear operator that maps a subspace of C(Ω) into C(Ω). Suppose that L reproduces polynomials up to degree m. We show that for appropriately defined coefficients amrj the operator $$H_{mr}[f]({\bf x}):= L \left[\sum\limits_{j=0}^{r} \frac{a_{mrj}}{j!} D^{j}_{{\bf x}-\cdot}\,f \right] ({\bf x}) \qquad ({\bf x} \in \Omega)$$ reproduces polynomials up to degree m+r. This is an immediate consequence of the main result (Theorem 3.1) which provides an integral representation of the error f(x) − Hmr[f](x). Special emphasis is given to positive linear operators L. In this case, sharp error bounds are established (Theorem 4.4) and interpolation properties are pointed out (Theorem 4.5). We also discuss various classes of admissible operators L and show an interrelation (Theorem 5.1).

Strict monotonicity of the spectral radii of bounded, positive, ordered linear operators is investigated. It is well-known that under reasonable assumptions, the spectral radii of two ordered positive operators enjoy a non-strict inequality. It is also well-known that a “strict” inequality between operators does not imply strict monotonicity of the spectral radii in general—some additional structure is required. We present a number of sufficient conditions on both the cone and the operators for such a strict ordering to hold which generalise known results in the literature, and have utility in comparison arguments, ubiquitous in positive systems theory.

The present paper deals with the A-statistical approximation processes of the general class of integral type linear positive operators including many well-known operators in the $$L_{p}$$ -metric spaces.

This paper deals with q-analogues of linear positive operators defined in weighted space of continuous functions defined on real axis. We study, using q-calculus, the existence of Korovkin's theorem in the spaces of continuous and unbounded functions defined on unbounded sets.

Let (X, F, μ) be a probability measure space, p and β real numbers such that 1≤p<+∞ and 0<β<p. For any linear positive operator T satisfying T1, T*1 = 1 we prove the norm and pointwise convergence of the sequence We get then the pointwise and norm convergence in Lp, 0 < β ≥ 1 < p < 2, of the sequence sgn Sif for any positive linear operator on Lp(Ω, A, μ) (μ-σ-finite) verifying ∥(1 − α)I + αS∥p ≤ 1 for a real number 0 < α < 1. In the particular case α = 1, (S is a contraction), β = p−l, this result gives the pointwise and norm convergence of the sequences introduced by Beauzamy and Enflo in 1985 to the asymptotic center of the sequence .

Let [a,b] be a compact interval of the real line IR. It is known that the classical theorem of Bohman-Korovkin states that in order that a sequence of positive linear operators (Lm), mapping into itself the space C[a,b] of continuous real-valued functions on [a,b], equipped with the uniform norm, to have the property that, for any f∈C[a,b], if m → ∞ we have lim Lmf = f, uniformly on [a,b], it is necessary and sufficient that such a convergence occur for a triplet of “test functions” from C[a,b], forming a so called Korovkin system. For C[a,b] the three monomials e0,e1,e2 where ej(x) := xj (j = 0,1,2), represent such a system.