1: \begin{abstract}

2: We study greedy approximation in uniformly smooth Banach spaces.

3: The Weak Chebyshev Greedy Algorithm (WCGA), introduced and 

4: studied in~\cite{t_gabs}, is defined for any Banach space $X$ and 

5: a dictionary $\D$, and provides nonlinear $n$-term approximation 

6: with respect to $\D$.

7: In this paper we study the Approximate Weak Chebyshev Greedy Algorithm 

8: (AWCGA)~-- a modification of the WCGA that was studied in~\cite{t_gtabsa}. 

9: In the AWCGA we are allowed to calculate $n$-term approximation with

10: a perturbation in computing the norming functional and 

11: a relative error in calculating the approximant. 

12: Such permission is natural for the numerical 

13: applications and simplifies realization of the algorithm.

14: We obtain conditions that are necessary and sufficient for the convergence of the 

15: AWCGA for any element of $X$. In particular, we show that 

16: if perturbations and errors are from $\ell_1$ space then the conditions 

17: for the convergence of the AWCGA are the same as for the WCGA.

18: For specifically chosen perturbations and errors we estimate the rate of 

19: convergence for any element $f$ from the closure of the convex hull of $\D$ 

20: and demonstrate that in special cases the AWCGA performs as well as the WCGA.

21: 

22: \smallskip\noindent

23: \textbf{Keywords:} Banach space; nonlinear approximation; 

24: greedy algorithm; Weak Chebyshev Greedy Algorithm; 

25: Approximate Weak Chebyshev Greedy Algorithm.

26: \end{abstract}

27: