No label defined (Q4778419)

Suppose that \(f(x)=(f_1(x), \ldots, f_r(x))^T, x\in \mathbb{R}^d\), is a vector-valued function satisfying the refinement equation \(f(x)=\sum_{k\in \mathbb{Z}^d} c_k f(2x-k)\), where \(\{c_k\}_{k\in \mathbb{Z}^d}\) is a finitely supported sequence of \(r\times r\) matrices. The approximation properties of the refinable function vector \(f\) have been extensively investigated in the literature in various ways. In this paper, under the assumption that the shifts of \(f\) are linearly independent, the authors characterize the accuracy order provided by the shift-invariant space generated by the function vector \(f\) in terms of the symbol of its mask. As the main result in this paper, Formulas (4.1) and (4.2) in Theorem 4.1 are useful in designing masks with good approximation properties and these two formulas are similar to the notion of the so-called sum rules of a matrix mask for a general dilation matrix, which has been discussed in [\textit{B. Han}, ``Approximation properties and construction of Hermite interpolants and biorthogonal multiwavelets'', J. Approximation Theory 110, 18-53 (2001; Zbl 0986.42020)].