• Title/Summary/Keyword: $G^2[C^1]$ Hermite interpolation

Search Result 2, Processing Time 0.025 seconds

HERMITE INTERPOLATION USING PH CURVES WITH UNDETERMINED JUNCTION POINTS

  • Kong, Jae-Hoon;Jeong, Seung-Pil;Kim, Gwang-Il
    • Bulletin of the Korean Mathematical Society
    • /
    • v.49 no.1
    • /
    • pp.175-195
    • /
    • 2012
  • Representing planar Pythagorean hodograph (PH) curves by the complex roots of their hodographs, we standardize Farouki's double cubic method to become the undetermined junction point (UJP) method, and then prove the generic existence of solutions for general $C^1$ Hermite interpolation problems. We also extend the UJP method to solve $C^2$ Hermite interpolation problems with multiple PH cubics, and also prove the generic existence of solutions which consist of triple PH cubics with $C^1$ junction points. Further generalizing the UJP method, we go on to solve $C^2$ Hermite interpolation problems using two PH quintics with a $C^1$ junction point, and we also show the possibility of applying the modi e UJP method to $G^2[C^1]$ Hermite interpolation.

HERMITE-TYPE EXPONENTIALLY FITTED INTERPOLATION FORMULAS USING THREE UNEQUALLY SPACED NODES

  • Kim, Kyung Joong
    • Communications of the Korean Mathematical Society
    • /
    • v.37 no.1
    • /
    • pp.303-326
    • /
    • 2022
  • Our aim is to construct Hermite-type exponentially fitted interpolation formulas that use not only the pointwise values of an 𝜔-dependent function f but also the values of its first derivative at three unequally spaced nodes. The function f is of the form, f(x) = g1(x) cos(𝜔x) + g2(x) sin(𝜔x), x ∈ [a, b], where g1 and g2 are smooth enough to be well approximated by polynomials. To achieve such an aim, we first present Hermite-type exponentially fitted interpolation formulas IN built on the foundation using N unequally spaced nodes. Then the coefficients of IN are determined by solving a linear system, and some of the properties of these coefficients are obtained. When N is 2 or 3, some results are obtained with respect to the determinant of the coefficient matrix of the linear system which is associated with IN. For N = 3, the errors for IN are approached theoretically and they are compared numerically with the errors for other interpolation formulas.