Skip to main content
2 of 6
added 37 characters in body
Michael Hardy
  • 1
  • 12
  • 85
  • 126

Defining Boundary Conditions for Spline Interpolation via the Euler–Maclaurin Formula

The Euler–Maclaurin formula states an interdependency between

\begin{align} I\quad:=&&\int_m^nf(x) \, dx;\ m,n\in\mathbb{Z}\\[6pt] S\quad:=&&\sum_{k=m}^n f(k) \\[6pt] D\quad:=&&\left\lbrace \frac{d^j}{dx^j} f(m) \right\rbrace \cup \left\lbrace \frac{d^j}{dx^j} f(n) \right\rbrace \end{align}


The Euler–Maclaurin formula is "traditionally" used to estimate the value of $S$ from $I$ and $D$ and also to estimate $I$ from $S$ and $D$

Question:

are there examples where the Euler–Maclaurin formula has been beneficially used to fix the the values of the derivatives of a function in the endpoints of its integration interval in order to fix the values of the elements of $D$ via $\lbrace I,\left(x_i,y_i\right)_{i=m}^n \rbrace$?

To be specific: have there been attempts to determine the values of the elements of $D$ via the Euler–Maclaurin formula to find "ideal" derivatives for use in clamped spline-interpolation, i.e. where the disambiguation of the interpolation is done via fixing derivatives at the ends of the interpolation interval by plugging the (estimated) value of $I$ and/or the coordinate values of the $\left(x_i,f(x_i)\right)$ into the Euler–Maclaurin formula?

Manfred Weis
  • 13.2k
  • 4
  • 34
  • 76