Difference between revisions of "Main Page"

From MWiki
Jump to: navigation, search
(Intex method)
(Representation theorems for integrals and derivatives)
(49 intermediate revisions by the same user not shown)
Line 1: Line 1:
 
__NOTOC__
 
__NOTOC__
 
= Welcome to MWiki =
 
= Welcome to MWiki =
−
== Theorem of the month ==
+
== Theorems of the month ==
−
The intex method solves every solvable LP in <math>\mathcal{O}({\vartheta}^{3})</math>.
+
=== Definition ===
  
−
== Proof and algorithm ==
+
Let <math>f_n^*(z) = f(\eta_nz)</math> <em>sisters</em> of the Taylor series <math>f(z) \in \mathcal{O}(\mathbb{D})</math> centred on 0 on the domain <math>\mathbb{D} \subseteq {}^{\omega}\mathbb{C}</math> where <math>m, n \in {}^{\omega}\mathbb{N}^{*}</math> and <math>\eta_n^m := \underline{1}^{2^{\lceil m/n \rceil}}</math>. Then let <math>\delta_n^*f = \tilde{2}(f - f_n^*)</math> the <em>halved sister distances</em> of <math>f.</math> For <math>\mu_n^m := m!n!/(m + n)!</math>, <math>\mu</math> and <math>\eta</math> form an calculus, which can be resolved on the level of Taylor series and allows an easy and finite closed representation of integrals and derivatives.<math>\triangle</math>
−
First, normalise and scale <math>{b}^{T}y - {c}^{T}x \le 0, Ax \le b</math> as well as <math>{A}^{T}y \ge c</math>. Let the <em>height</em> <math>h</math> have the initial value <math>h_0 := s |\min \; \{b_1, ..., b_m, -d_1, ..., -d_n\}|</math> for the <em>elongation factor</em> <math>s \in \, ]1, 2]</math>.</br>
 
−
The LP min <math>\{h \in [0, h_0] : x \in {}^{\omega}\mathbb{R}_{\ge 0}^{n}, y \in {}^{\omega}\mathbb{R}_{\ge 0}^{m},{b}^{T}y - {c}^{T}x \le h, Ax - b \le (h, ..., h)^T \in {}^{\omega}\mathbb{R}_{\ge 0}^{m}, c - {A}^{T}y \le (h, ..., h)^T \in {}^{\omega}\mathbb{R}_{\ge 0}^{n}\}</math> has <math>k</math> constraints and the feasible starting point <math>(x_0, y_0, h_0/s)^{T} \in {}^{\omega}\mathbb{R}_{\ge 0}^{m+n+1}</math>, e.g. <math>(0, 0, h_0/s)^{T}</math>.</br>
 
−
It identifies the mutually dual LPs max <math>\{{c}^{T}x : c \in {}^{\omega}\mathbb{R}^{n}, x \in {P}_{\ge 0}\}</math> and min <math>\{{b}^{T}y : y \in {}^{\omega}\mathbb{R}_{\ge 0}^{m}, {A}^{T}y \ge c\}</math>.
 
  
−
Let the point <math>p := (x, y, h)^T</math> approximate the subpolytope’s centre of gravity <math>P^*</math> as <math>p_k^* := (\min p_k + \max p_k)/2</math> until <math>{|| \Delta p ||}_{1}</math> is sufficiently small. Here <math>x</math> takes precedence over <math>y</math>. Then extrapolate <math>p</math> via <math>{p}^{*}</math> into <math>\partial P^*</math> as <math>u</math>. Put <math>p := p^* + (u - p^*)/s</math> to shun <math>\partial P^*</math>. Hereon approximate <math>p</math> more deeply again as centre of gravity. After optionally solving all LPs min<math>{}_{k} {h}_{k}</math> by bisection methods for <math>{h}_{k} \in {}^{\omega}\mathbb{R}_{\ge 0}</math> in <math>\mathcal{O}({\vartheta}^{2})</math> each time, <math>v \in {}^{\omega}\mathbb{R}^{k}</math> may be determined such that <math>v_k := \Delta{p}_{k} \Delta{h}_{k}/r</math> and <math>r :=</math> min<math>{}_{k} \Delta{h}_{k}</math>. Simplified let <math>|\Delta{p}_{1}| = ... = |\Delta{p}_{m+n}|</math>.
+
=== Representation theorem for integrals ===
  
−
Here min <math>{h}_{m+n+1}</math> may be solved for <math>p^* := p + tv</math> where <math>t \in {}^{\omega}\mathbb{R}_{\ge 0}</math> and <math>{v}_{m+n+1} = 0</math>. If min<math>{}_{k} {h}_{k} r = 0</math> follows, end, otherwise start over until min <math>h = 0</math> or min <math>h &gt; 0</math> is certain. If necessary, relax the constraints temporarily by the same small modulus.</br>
+
The Taylor series (see below) <math>f(z) \in \mathcal{O}(\mathbb{D})</math> centred on 0 on <math>\mathbb{D} \subseteq {}^{\omega}\mathbb{C}</math> gives for <math>\grave{m}, n \in {}^{\omega}\mathbb{N}^*</math><div style="text-align:center;"><math>{\uparrow}_0^z...{\uparrow}_0^{\zeta_2}{f(\zeta_1){\downarrow}\zeta_1\;...\;{\downarrow}\zeta_n} = \widetilde{n!} f(z\mu_n) z^n.\square</math></div>
−
Since almost every iteration step in <math>\mathcal{O}({\omega\vartheta}^{2})</math> halves <math>h</math> at least, the strong duality theorem yields the result.<math>\square</math>
+
 
 +
=== Representation theorem for derivatives ===
 +
 
 +
For <math>{}^{\widetilde{\nu}}\dot{\mathbb{C}} \subset \mathbb{D} \subseteq {}^{\omega}\mathbb{C},</math> the Taylor series<div style="text-align:center;"><math>f(z):=f(0) + {\LARGE{\textbf{+}}}_{m=1}^{\omega }{\widetilde{m!}\,{{f}^{(m)}}(0){z^m}},</math></div><math>\varepsilon := \tilde{2}^j\tilde{r}, j \in {}^{\omega}\mathbb{Z}, n = \epsilon^{\sigma} \in {}^{\omega}\mathbb{N}^{*}, u :=\epsilon^{\tilde{n} \hat{\underline{\pi}}}</math> and <math>f</math>'s radius of convergence <math>r \in {}^{\nu}{\mathbb{R}}_{&gt;0}</math> imply<div style="text-align:center;"><math>{{f}^{(n)}}(0)=2^{jn}\acute{n}!{\LARGE{\textbf{+}}}_{k=1}^{n}{\delta_n^* f(\tilde{2}^j u^k)}.</math></div>
 +
 
 +
==== Proof: ====
 +
Taylor's theorem<ref name="Remmert">[[w:Reinhold Remmert|<span class="wikipedia">Remmert, Reinhold</span>]]: ''Funktionentheorie 1'' : 3., verb. Aufl.; 1992; Springer; Berlin; ISBN 9783540552338, S. 165 f.</ref> and the properties of the roots of unity.<math>\square</math>
 +
 
 +
== Reference ==
 +
<references />
 +
 
 +
== Recommended reading ==
  
−
== Recommended readings ==
 
 
[https://en.calameo.com/books/003777977258f7b4aa332 Nonstandard Mathematics]
 
[https://en.calameo.com/books/003777977258f7b4aa332 Nonstandard Mathematics]
  
 
[[de:Hauptseite]]
 
[[de:Hauptseite]]

Revision as of 23:05, 31 March 2024

Welcome to MWiki

Theorems of the month

Definition

Let [math]\displaystyle{ f_n^*(z) = f(\eta_nz) }[/math] sisters of the Taylor series [math]\displaystyle{ f(z) \in \mathcal{O}(\mathbb{D}) }[/math] centred on 0 on the domain [math]\displaystyle{ \mathbb{D} \subseteq {}^{\omega}\mathbb{C} }[/math] where [math]\displaystyle{ m, n \in {}^{\omega}\mathbb{N}^{*} }[/math] and [math]\displaystyle{ \eta_n^m := \underline{1}^{2^{\lceil m/n \rceil}} }[/math]. Then let [math]\displaystyle{ \delta_n^*f = \tilde{2}(f - f_n^*) }[/math] the halved sister distances of [math]\displaystyle{ f. }[/math] For [math]\displaystyle{ \mu_n^m := m!n!/(m + n)! }[/math], [math]\displaystyle{ \mu }[/math] and [math]\displaystyle{ \eta }[/math] form an calculus, which can be resolved on the level of Taylor series and allows an easy and finite closed representation of integrals and derivatives.[math]\displaystyle{ \triangle }[/math]

Representation theorem for integrals

The Taylor series (see below) [math]\displaystyle{ f(z) \in \mathcal{O}(\mathbb{D}) }[/math] centred on 0 on [math]\displaystyle{ \mathbb{D} \subseteq {}^{\omega}\mathbb{C} }[/math] gives for [math]\displaystyle{ \grave{m}, n \in {}^{\omega}\mathbb{N}^* }[/math]

[math]\displaystyle{ {\uparrow}_0^z...{\uparrow}_0^{\zeta_2}{f(\zeta_1){\downarrow}\zeta_1\;...\;{\downarrow}\zeta_n} = \widetilde{n!} f(z\mu_n) z^n.\square }[/math]

Representation theorem for derivatives

For [math]\displaystyle{ {}^{\widetilde{\nu}}\dot{\mathbb{C}} \subset \mathbb{D} \subseteq {}^{\omega}\mathbb{C}, }[/math] the Taylor series

[math]\displaystyle{ f(z):=f(0) + {\LARGE{\textbf{+}}}_{m=1}^{\omega }{\widetilde{m!}\,{{f}^{(m)}}(0){z^m}}, }[/math]

[math]\displaystyle{ \varepsilon := \tilde{2}^j\tilde{r}, j \in {}^{\omega}\mathbb{Z}, n = \epsilon^{\sigma} \in {}^{\omega}\mathbb{N}^{*}, u :=\epsilon^{\tilde{n} \hat{\underline{\pi}}} }[/math] and [math]\displaystyle{ f }[/math]'s radius of convergence [math]\displaystyle{ r \in {}^{\nu}{\mathbb{R}}_{>0} }[/math] imply

[math]\displaystyle{ {{f}^{(n)}}(0)=2^{jn}\acute{n}!{\LARGE{\textbf{+}}}_{k=1}^{n}{\delta_n^* f(\tilde{2}^j u^k)}. }[/math]

Proof:

Taylor's theorem[1] and the properties of the roots of unity.[math]\displaystyle{ \square }[/math]

Reference

  1. ↑ Remmert, Reinhold: Funktionentheorie 1 : 3., verb. Aufl.; 1992; Springer; Berlin; ISBN 9783540552338, S. 165 f.

Recommended reading

Nonstandard Mathematics

Retrieved from "https://en.mwiki.de/index.php?title=Main_Page&oldid=298"