History of Nodes and weights of Gauss–Kronrod quadrature

back to table · edit · history · where entries came from · files

compare when who what
2026-09-10 09:20 bmatschke interactive the differential operator is upright; d is not a variable current reviewed
2026-09-06 02:42 bmatschke interactive the field is named repeats: the numberdb package already uses restating for rewriting an entry in different digits, which is a different thing, and its argument is published
2026-09-05 19:06 bmatschke interactive says which table states the shared values first: the nodes and weights of the embedded Gauss rule are the Gauss-Legendre nodes and weights, computed the same way. Search folds the repeat into the original rather than answering the same number twice
2026-09-05 11:32 bmatschke interactive the completeness note is read as part of a sentence and was a noun phrase with a telegraphic appendix; say which rules are here, what n indexes, and that each is listed in full
2026-09-05 11:27 bmatschke interactive the completeness note is read as part of a sentence and was a noun phrase with a telegraphic appendix; say which rules are here, what n indexes, and that each is listed in full
2026-09-03 02:51 bmatschke interactive define the Stieltjes polynomial in Formulas, where the first formula uses it and the page draws it before the comment that defined it; drop a comment that restates formula (4) and ends on what a search returns; and say that a hundred digits are written rather than as many as the ball supports -- all
2026-09-03 02:06 zeta3 with claude (agent run 20260903T012415Z) table-build@1.42+4513583d audit: a shorter Definition; what the nodes are is in the first comment
2026-09-03 01:49 zeta3 with claude (agent run 20260903T0124 table-build@1.42+4513583d Gauss-Kronrod nodes and weights for n <= 15 and n = 20, 25, 30: Stieltjes polynomials solved exactly over Q, roots isolated over Q[x], weights in ball arithmetic, every rule checked for its degree of exactness with a control before being sent
2026-09-03 01:49 zeta3 table-build@1.42+4513583d checking that this table can be written to
2026-09-03 01:49 zeta3 with claude (agent run 20260903T012415Z) table-build@1.42+4513583d draft: nodes and weights of Gauss-Kronrod quadrature, proposal 4 of BATCH-2026-09-02

What changed between 2026-09-06 02:42 and 2026-09-10 09:20

from line 1 (6 lines) @@ -1,6 +1,6 @@
 Title: Nodes and weights of Gauss–Kronrod quadrature Definition: For $n\geq 1$ the Gauss–Kronrod rule with $2n+1$ points, $\int_{-1}^{1}-  f(x)\,dx \approx \sum_{k=1}^{2n+1} w_k f(x_k)$, extends the Gauss–Legendre rule-  with $n$ points by $n+1$ nodes so as to be exact for every polynomial of degree+  f(x)\,\mathrm{d}x \approx \sum_{k=1}^{2n+1} w_k f(x_k)$, extends the Gauss–Legendre+  rule with $n$ points by $n+1$ nodes so as to be exact for every polynomial of degree   at most $3n+1$ CITE{Wiki}. Listed are its nodes $x_1<\cdots<x_{2n+1}$ and its weights   $w_k$.
from line 62 (12 lines, 1 more than before) @@ -62,11 +62,12 @@
     the HREF{Legendre_polynomials}[Legendre polynomial] of degree $n$ and $E_{n+1}$     is the Stieltjes polynomial CITE{Wiki-Stieltjes}: the monic polynomial of degree-    $n+1$ with $\int_{-1}^{1}P_n(x)E_{n+1}(x)\,x^j\,dx=0$ for $0\leq j\leq n$.'+    $n+1$ with $\int_{-1}^{1}P_n(x)E_{n+1}(x)\,x^j\,\mathrm{d}x=0$ for $0\leq j\leq+    n$.'   formula-stieltjes: $E_2=x^2-\tfrac35$, $E_3=x^3-\tfrac67x$, $E_4=x^4-\tfrac{10}{9}x^2+\tfrac{155}{891}$,     $E_5=x^5-\tfrac{15}{11}x^3+\tfrac{615}{1573}x$, $E_6=x^6-\tfrac{21}{13}x^4+\tfrac{567}{845}x^2-\tfrac{8043}{186745}$.     $E_{n+1}$ has the parity of $n+1$, so $x=0$ is a root of $E_{n+1}$ for even $n$     and of $P_n$ for odd $n$.-  formula-weights: $w_k=\dfrac{q_{\omega_n}(x_k)}{\omega_n'(x_k)}$ with $q_{\omega_n}(x)=\int_{-1}^{1}\dfrac{\omega_n(t)-\omega_n(x)}{t-x}\,dt$,-    and $q_{\omega_n}=c_n+E_{n+1}\,q_n$, where $c_n=\int_{-1}^{1}P_n(x)\,x^n\,dx=\dfrac{2^{n+1}(n!)^2}{(2n+1)!}$+  formula-weights: $w_k=\dfrac{q_{\omega_n}(x_k)}{\omega_n'(x_k)}$ with $q_{\omega_n}(x)=\int_{-1}^{1}\dfrac{\omega_n(t)-\omega_n(x)}{t-x}\,\mathrm{d}t$,+    and $q_{\omega_n}=c_n+E_{n+1}\,q_n$, where $c_n=\int_{-1}^{1}P_n(x)\,x^n\,\mathrm{d}x=\dfrac{2^{n+1}(n!)^2}{(2n+1)!}$     and $q_n$ is the HREF{Secondary_polynomials_of_the_Legendre_polynomials}[secondary     polynomial of $P_n$]. Hence at a Gauss node $w_{2j}=\lambda_j+\dfrac{c_n}{P_n'(x_{2j})\,E_{n+1}(x_{2j})}$
from line 76 (5 lines) @@ -75,5 +76,5 @@
     to $\lambda_j$ is negative because $P_n'$ and $E_{n+1}$ take opposite signs at     every root of $P_n$, by the interlacing.-  formula-exactness: $\sum_{k=1}^{2n+1} w_k x_k^m=\int_{-1}^{1}x^m\,dx=\dfrac{1+(-1)^m}{m+1}$+  formula-exactness: $\sum_{k=1}^{2n+1} w_k x_k^m=\int_{-1}^{1}x^m\,\mathrm{d}x=\dfrac{1+(-1)^m}{m+1}$     for $0\leq m\leq 3n+1$ CITE{Laurie}, and by the symmetry also for $m=3n+2$ when     $n$ is odd; it fails at the next even $m$. So the degree of exactness is $3n+1$
from line 84 (5 lines) @@ -83,5 +84,5 @@
     is rational: $\tfrac89$, $\tfrac{28}{45}$, $\tfrac{22016}{48825}$, $\tfrac{201344}{581175}$,     $\tfrac{118298624}{418034925}$ for $n=1,\ldots,5$.'-  formula-lagrange: $w_k=\int_{-1}^{1}\ell_k(x)\,dx$ with $\ell_k(x)=\prod_{j\neq+  formula-lagrange: $w_k=\int_{-1}^{1}\ell_k(x)\,\mathrm{d}x$ with $\ell_k(x)=\prod_{j\neq     k}\frac{x-x_j}{x_k-x_j}$ the Lagrange basis polynomial of the $2n+1$ nodes, as     for every interpolatory rule; on HREF{Lagrange_basis_polynomials_for_equally_spaced_nodes}[equally 

Sign in to restore an earlier version.