Vector subdivision schemes in ( L p (ℝ s )) r (1⩽ p ⩽∞) spaces

Song Li

DOI: https://doi.org/10.1360/03ys9039

2003-01-01

Abstract:The purpose of this paper is to investigate the refinement equations of the form φ (x) = ∑_α∈ℤ^s a(α )φ (Mx - α ), x ∈ℝ^s , where the vector of functions ϕ =( ϕ 1 ..., ϕ r ) T is in ( L p (ℝ s )) r , 1⩽ p ⩽∞, a (α), α∈ℤ s is a finitely supported sequence of r × r matrices called the refinement mask, and M is an s × s integer matrix such that lim→∞ M -n = 0. In order to solve the refinement equation mentioned above, we start with a vector of compactly supported functions φ 0 ∈( L p (ℝ s )) r and use the iteration schemes f n := Q a n φ 0 , n =1,2,..., where Q n is the linear operator defined on ( L p (ℝ s )) r given by Q_a φ : = ∑_α∈ℤ^s a(α )φ (M · - α ), φ∈ (L_p (ℝ^s ))^r . This iteration scheme is called a subdivision scheme or cascade algorithm. In this paper, we characterize the L p -convergence of subdivision schemes in terms of the p-norm joint spectral radius of a finite collection of some linear operators determined by the sequence a and the set B restricted to a certain invariant subspace, where the set B is a complete set of representatives of the distinct cosets of the quotient group ℤ s /Mℤ s containing 0.