A Parallel Algorithm for Interval Polynomial Interpolation

Chenyi Hu
Abstract:AbstractBy taking data measurement errors into considerations, numerical results obtainedby interval polynomial interpolation [5] are highly reliable. To reduce overestimation,combinations of interval nodes are used. The high computational complexity makes itnecessary to perform the computation parallelly. In this paper, we report a parallelalgorithm which e ectively perform the combinatorical computation.A brief introduction of interval polynomial interpolation is reviewed in section 1.Currently used sequential algorithms are reported in section 2. A parallel algorithmwith discussions on computation decomposition, mapping, and load balancing ispresented in section 3. A numerical example is presented in section 4. We concludethis paper with section 5. 1 Introduction Interval computations are used in this paper. Throughout this paper, boldfaced letters suchas x ; y represent interval variables while x;y represent real numbers. For an interval x,weuse x and x to represent its greatest lower bound and least upper bound respectively, i.e.x =[
What problem does this paper attempt to address?