Interpolating refinable function vectors with reflection properties (Q893132)

A refinable function vector \(\Phi=(\phi_1,\ldots,\phi_r)^T\) satisfies an refinement equation: \[ \Phi(x)=N\sum_{k\in \mathbb{Z}} P_k \Phi(Nx-k) \] for a finitely supported sequence \(\{P_k\}_{k\in \mathbb{Z}}\) and a positive integer \(N>1\). In this paper, the authors consider a special family of refinable function vectors \(\Phi\) having the interpolation property and the (two-direction) reflection property, i.e., (a) the functions \(\phi_1,\ldots,\phi_r\) satisfy the interpolation condition: \(\phi_j(n+\frac{m-1}{r})=\delta_{n,0} \delta_{j,m}\) for all \(j,m=1,\ldots,r\) and \(n\in \mathbb{Z}\); (b) the functions \(\phi_1,\ldots,\phi_r\) satisfy the reflection condition: \(\phi_j(x)=\phi_{r-j+1}(2\alpha-x)\) for all \(j=1,\ldots,r\) for some \(\alpha\in \mathbb{R}\). The authors first study the approximation property and polynomial reproduction property of a general refinable function vector \(\Phi\). Then the authors characterize in Theorem 3.3 such refinable function vector, which has the interpolation property and reflection property. Several detailed examples of such refinable function vectors with interpolation and reflection properties are provided in Section~4. The constructed refinable function vectors are of interest in numerical mathematics and wavelet analysis.