Solutions in Sobolev spaces of vector refinement equations with a general dilation matrix (Q2491228)

An \(s\times s\) integer matrix \(M\) is called a dilation matrix if all its eigenvalues are greater than 1 in modulus. In this interesting paper, the author considers the vector refinement equation \[ \phi = | \det M| \, \sum_{\beta\in {\mathbb Z}^s}a(\beta)\,\phi(M\cdot - \beta) \] where \(\phi=(\phi_1,\ldots,\phi_r)^{\text{T}}\) is a vector of compactly supported functions or distributions, and \(a\) is a finitely supported matrix mask. He presents a necessary and sufficient condition for the existence of solutions \(\phi\) in a Sobolev space \(W_p^k({\mathbb R}^s)\). This criterion is constructive. The rate of convergence of the cascade algorithms in \(W_p^k({\mathbb R}^s)\) is investigated. If all eigenvalues of \(M\) have the same modulus, the author presents a characterization for the \(L_p\) critical smoothness exponent of an \(M\)-refinable function vector without the assumption of stability on the refinable function vector. This paper generalizes results in \textit{R. Q. Jia}, \textit{K. S. Lau} and \textit{D.X. Zhou} [J. Fourier Anal. Appl. 7, 143--167 (2001; Zbl 1030.42031)] in the univariate setting to the multivariate case.