On the Bézier variant of generalized Kantorovich type Balazs operators (Q812657)

From MaRDI portal
Revision as of 18:26, 19 March 2024 by Maintenance script (talk | contribs) (rollbackEdits.php mass rollback)
scientific article
Language Label Description Also known as
English
On the Bézier variant of generalized Kantorovich type Balazs operators
scientific article

    Statements

    On the Bézier variant of generalized Kantorovich type Balazs operators (English)
    0 references
    0 references
    24 January 2006
    0 references
    The authors introduce a Kantorovich-type operator defined by \[ L^*_n(f;x)=na_n\sum^\infty_{k=0} p_{n,k}(x)\int_{I_{n,k}}f(t)\,dt,\quad x\in [0,\infty), \] where \(a_n\) are suitable chosen positive numbers independent of \(x\), \(I_{n,k}=[k/na_n,(k+1)/na_n]\) and \(p_{n,k}(x)= \frac{\phi^{(k)}_n(0)}{k!}\cdot \frac{(a_nx)^k}{\phi_n(a_nx)}\) with \(\{\phi_n\}\) a sequence of functions \(\phi_n:\mathbb C\to\mathbb C\) satisfying the following conditions: (i) \(\phi_n\) \((n=1,2,\dots)\) is analytic in a domain \(D\) containing the disk \(B=\{z\in \mathbb C: |z-b|\leq b\}\); (ii) \(\phi_n(0)=1\) \((n=1,2,\dots)\); (iii) for any \(x\geq 0\), \(\phi_n(x)>0\) and \(\phi_n^{(k)}(0)\geq 0\) for any \(n=1,2,\dots\) and \(k=1,2,\dots\); (iv) for every \(n=1,2,\dots,\) \(\frac{\phi^{(v)}_n (a_nx)}{n^v\phi_n(a_nx)}= 1+O(\frac1{na_n})\), \(v=1,2,3,4,\) where \(a_n\to 0\), \(na_n\to\infty\) as \(n\to\infty\). Moreover the authors propose the Bézier variant of \(L^*_n\) as \[ L^*_{n,\alpha}(f;x)=na_n\sum^\infty_{k=0} Q^{(\alpha)}_{n,k}(x)\int_{n,k}f(t)\,dt,\quad x\in[0,\infty), \] where \(Q^{(\alpha)}_{n,k}(x)=\{\sum^\infty_{j=k} p_{n,j}(x)\}^\alpha-\{\sum^\infty_{j=k+1}p_{n,j}(x)\}^\alpha,\quad \alpha\geq 1.\) In the main result is estimated the rate of convergence by \(L^*_{n,\alpha}\) for functions of bounded variation as follows: Theorem. Let \(f\) be a function of bounded variation on every finite subinterval of \([0,\infty)\) and \(\mathop V^b_a(g_x)\) is the total variation of \(g_x\) on \([a,b]\). If \(\alpha\geq 1\), \(x\in (0,\infty)\), \(r>1\) and \(M>0\) are given and \(f(t)=O(t^r)\), \(t\to \infty\), then for \(n\) sufficiently large, \[ \begin{multlined} \left| L^*_{n,\alpha}(f,x)-\frac{1}{2^a}\,f(x+)-\left(1-\frac{1}{2^\alpha}\right)f(x-)\right|\leq A(n,x)| f(x+)-f(x-)|+\\+ \frac{2\alpha M(1+x+x^2)+a_nx^2}{na_nx^2}\;\sum^2_{k=1}V^{x+x/\sqrt k}_{x-x/\sqrt k}(g_x)+O(n^{-r}),\end{multlined} \] where \[ A(n,x)\leq\begin{cases} \frac{\alpha 2^{\alpha-1}[6\sqrt{1+x}+1]}{\sqrt{nx}}\,,\quad & a_n=1,\;\phi_n(x)=e^{nx}\\ a2^\alpha \left[\frac{[1+(a_nx)^2+ 0.5(1+a_nx)^2]}{(1+a_nx)[1+\sqrt na_nx]}+\frac{1+a_nx}{\sqrt {2ena_nx}}\right],\quad & \phi_n(x)=(1+x)^n\end{cases} \] and \[ g_x(t)=\begin{cases} f(t)-f(x-),\quad & 0\leq t<x\\ 0,\quad & t=x\\ f(t)-f(x+),\quad & x<t<\infty.\end{cases} \]
    0 references
    0 references
    Balázs operators
    0 references
    rate of convergence
    0 references
    bounded variation
    0 references
    total variation
    0 references
    Bézier basis
    0 references

    Identifiers