\documentclass[11 pt,xcolor={dvipsnames,svgnames,x11names,table}]{beamer} \usepackage[english]{babel} \usepackage{polski} \usetheme[ bullet=circle, % Other option: square bigpagenumber, % circled page number on lower right topline=true, % colored bar at the top of the frame shadow=false, % Shading for beamer blocks watermark=BG_lower, % png file for the watermark ]{Flip} %\logo{\kern+1.em\includegraphics[height=1cm]{SHiP-3_LightCharcoal}} \usepackage[lf]{berenis} \usepackage[LY1]{fontenc} \usepackage[utf8]{inputenc} \usepackage{cases} \usepackage{mathtools} \usepackage{emerald} \usefonttheme{professionalfonts} \usepackage[no-math]{fontspec} \defaultfontfeatures{Mapping=tex-text} % This seems to be important for mapping glyphs properly \setmainfont{Gillius ADF} % Beamer ignores "main font" in favor of sans font \setsansfont{Gillius ADF} % This is the font that beamer will use by default % \setmainfont{Gill Sans Light} % Prettier, but harder to read \setbeamerfont{title}{family=\fontspec{Gillius ADF}} \input t1augie.fd %\newcommand{\handwriting}{\fontspec{augie}} % From Emerald City, free font %\newcommand{\handwriting}{\usefont{T1}{fau}{m}{n}} % From Emerald City, free font % \newcommand{\handwriting}{} % If you prefer no special handwriting font or don't have augie %% Gill Sans doesn't look very nice when boldfaced %% This is a hack to use Helvetica instead %% Usage: \textbf{\forbold some stuff} %\newcommand{\forbold}{\fontspec{Arial}} \usepackage{graphicx} \usepackage[export]{adjustbox} \usepackage{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{bm} \usepackage{colortbl} \usepackage{mathrsfs} % For Weinberg-esque letters \usepackage{cancel} % For "SUSY-breaking" symbol \usepackage{slashed} % for slashed characters in math mode \usepackage{bbm} % for \mathbbm{1} (unit matrix) \usepackage{amsthm} % For theorem environment \usepackage{multirow} % For multi row cells in table \usepackage{arydshln} % For dashed lines in arrays and tables \usepackage{siunitx} \usepackage{xhfill} \usepackage{grffile} \usepackage{textpos} \usepackage{subfigure} \usepackage{tikz} \usepackage{hyperref} %\usepackage{hepparticles} \usepackage[italic]{hepparticles} \usepackage{hepnicenames} % Drawing a line \tikzstyle{lw} = [line width=20pt] \newcommand{\topline}{% \tikz[remember picture,overlay] {% \draw[crimsonred] ([yshift=-23.5pt]current page.north west) -- ([yshift=-23.5pt,xshift=\paperwidth]current page.north west);}} % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % \usepackage{tikzfeynman} % For Feynman diagrams \usetikzlibrary{arrows,shapes} \usetikzlibrary{trees} \usetikzlibrary{matrix,arrows} % For commutative diagram % http://www.felixl.de/commu.pdf \usetikzlibrary{positioning} % For "above of=" commands \usetikzlibrary{calc,through} % For coordinates \usetikzlibrary{decorations.pathreplacing} % For curly braces % http://www.math.ucla.edu/~getreuer/tikz.html \usepackage{pgffor} % For repeating patterns \usetikzlibrary{decorations.pathmorphing} % For Feynman Diagrams \usetikzlibrary{decorations.markings} \tikzset{ % >=stealth', %% Uncomment for more conventional arrows vector/.style={decorate, decoration={snake}, draw}, provector/.style={decorate, decoration={snake,amplitude=2.5pt}, draw}, antivector/.style={decorate, decoration={snake,amplitude=-2.5pt}, draw}, fermion/.style={draw=gray, postaction={decorate}, decoration={markings,mark=at position .55 with {\arrow[draw=gray]{>}}}}, fermionbar/.style={draw=gray, postaction={decorate}, decoration={markings,mark=at position .55 with {\arrow[draw=gray]{<}}}}, fermionnoarrow/.style={draw=gray}, gluon/.style={decorate, draw=black, decoration={coil,amplitude=4pt, segment length=5pt}}, scalar/.style={dashed,draw=black, postaction={decorate}, decoration={markings,mark=at position .55 with {\arrow[draw=black]{>}}}}, scalarbar/.style={dashed,draw=black, postaction={decorate}, decoration={markings,mark=at position .55 with {\arrow[draw=black]{<}}}}, scalarnoarrow/.style={dashed,draw=black}, electron/.style={draw=black, postaction={decorate}, decoration={markings,mark=at position .55 with {\arrow[draw=black]{>}}}}, bigvector/.style={decorate, decoration={snake,amplitude=4pt}, draw}, } % TIKZ - for block diagrams, % from http://www.texample.net/tikz/examples/control-system-principles/ % \usetikzlibrary{shapes,arrows} \tikzstyle{block} = [draw, rectangle, minimum height=3em, minimum width=6em] \usetikzlibrary{backgrounds} \usetikzlibrary{mindmap,trees} % For mind map \newcommand{\degree}{\ensuremath{^\circ}} \newcommand{\E}{\mathrm{E}} \newcommand{\Var}{\mathrm{Var}} \newcommand{\Cov}{\mathrm{Cov}} \newcommand\Ts{\rule{0pt}{2.6ex}} % Top strut \newcommand\Bs{\rule[-1.2ex]{0pt}{0pt}} % Bottom strut \graphicspath{{images/}} % Put all images in this directory. Avoids clutter. % SOME COMMANDS THAT I FIND HANDY % \renewcommand{\tilde}{\widetilde} % dinky tildes look silly, dosn't work with fontspec \newcommand{\comment}[1]{\textcolor{comment}{\footnotesize{#1}\normalsize}} % comment mild \newcommand{\Comment}[1]{\textcolor{Comment}{\footnotesize{#1}\normalsize}} % comment bold \newcommand{\COMMENT}[1]{\textcolor{COMMENT}{\footnotesize{#1}\normalsize}} % comment crazy bold \newcommand{\Alert}[1]{\textcolor{Alert}{#1}} % louder alert \newcommand{\ALERT}[1]{\textcolor{ALERT}{#1}} % loudest alert %% "\alert" is already a beamer pre-defined \newcommand*{\Scale}[2][4]{\scalebox{#1}{$#2$}}% \def\Put(#1,#2)#3{\leavevmode\makebox(0,0){\put(#1,#2){#3}}} \usepackage{gmp} \usepackage[final]{feynmp-auto} \usepackage[backend=bibtex,style=numeric-comp,firstinits=true]{biblatex} \bibliography{bib} \setbeamertemplate{bibliography item}[text] \makeatletter\let\frametextheight\beamer@frametextheight\makeatother % suppress frame numbering for backup slides % you always need the appendix for this! \newcommand{\backupbegin}{ \newcounter{framenumberappendix} \setcounter{framenumberappendix}{\value{framenumber}} } \newcommand{\backupend}{ \addtocounter{framenumberappendix}{-\value{framenumber}} \addtocounter{framenumber}{\value{framenumberappendix}} } \definecolor{links}{HTML}{2A1B81} %\hypersetup{colorlinks,linkcolor=,urlcolor=links} % For shapo's formulas: \def\lsi{\raise0.3ex\hbox{$<$\kern-0.75em\raise-1.1ex\hbox{$\sim$}}} \def\gsi{\raise0.3ex\hbox{$>$\kern-0.75em\raise-1.1ex\hbox{$\sim$}}} \newcommand{\lsim}{\mathop{\lsi}} \newcommand{\gsim}{\mathop{\gsi}} \newcommand{\wt}{\widetilde} %\newcommand{\ol}{\overline} \newcommand{\Tr}{\rm{Tr}} \newcommand{\tr}{\rm{tr}} \newcommand{\eqn}[1]{&\hspace{-0.7em}#1\hspace{-0.7em}&} \newcommand{\vev}[1]{\rm{$\langle #1 \rangle$}} \newcommand{\abs}[1]{\rm{$\left| #1 \right|$}} \newcommand{\eV}{\rm{eV}} \newcommand{\keV}{\rm{keV}} \newcommand{\GeV}{\rm{GeV}} \newcommand{\im}{\rm{Im}} \newcommand{\disp}{\displaystyle} \def\be{\begin{equation}} \def\ee{\end{equation}} \def\ba{\begin{eqnarray}} \def\ea{\end{eqnarray}} \def\d{\partial} \def\l{\left(} \def\r{\right)} \def\la{\langle} \def\ra{\rangle} \def\e{{\rm e}} \def\Br{{\rm Br}} \def\fixme{{\color{red} FIXME!}} \def\mc{{\color{Magenta}{MC}}\xspace} \def\pdf{{\rm p.d.f.}} \def\ARROW{{\color{JungleGreen}{$\Rrightarrow$}}\xspace} \def\ARROWR{{\color{WildStrawberry}{$\Rrightarrow$}}\xspace} \author{ {\fontspec{Trebuchet MS}Marcin Chrz\k{a}szcz, Danny van Dyk} (UZH)} \institute{UZH} \title[Linear equation systems: exact methods]{Linear equation systems:exact methods} \date{21 September 2016} \begin{document} \tikzstyle{every picture}+=[remember picture] { \setbeamertemplate{sidebar right}{\llap{\includegraphics[width=\paperwidth,height=\paperheight]{bubble2}}} \begin{frame}[c]%{\phantom{title page}} \begin{center} \begin{center} \begin{columns} \begin{column}{0.9\textwidth} \flushright\fontspec{Trebuchet MS}\bfseries \Huge {Linear equation systems: exact methods} \end{column} \begin{column}{0.2\textwidth} %\includegraphics[width=\textwidth]{SHiP-2} \end{column} \end{columns} \end{center} \quad \vspace{3em} \begin{columns} \begin{column}{0.6\textwidth} \flushright \vspace{-1.8em} {\fontspec{Trebuchet MS} \large Marcin ChrzÄ…szcz, Danny van Dyk\\\vspace{-0.1em}\small \href{mailto:mchrzasz@cern.ch}{mchrzasz@cern.ch}, \href{mailto:dany.van.dyk@gmail.com}{danny.van.dyk@gmail.com}} \end{column} \begin{column}{0.4\textwidth} \includegraphics[height=1.3cm]{uzh-transp} \end{column} \end{columns} \vspace{1em} % \footnotesize\textcolor{gray}{With N. Serra, B. Storaci\\Thanks to the theory support from M. Shaposhnikov, D. Gorbunov}\normalsize\\ \vspace{0.5em} \textcolor{normal text.fg!50!Comment}{Numerical Methods, \\ 10 October, 2016} \end{center} \end{frame} } \begin{frame}\frametitle{Linear eq. system} \ARROW This and the next lecture will focus on a well known problem. Solve the following equation system: \begin{align*} A \cdot x =b, \end{align*} \ARROWR $A = a_{ij} \in \mathbb{R}^{n\times n}$ and $\det(A) \neq 0$\\ \ARROWR $b=b_i \in \mathbb{R}^n$.\\ \ARROW The problem: Find the $x$ vector. \end{frame} \begin{frame}\frametitle{Error digression} \begin{small} \ARROW There is enormous amount of ways to solve the linear equation system.\\ \ARROW The choice of one over the other of them should be gathered by the {\it condition} of the matrix $A$ denoted at $cond(A)$. \ARROW If the $cond(A)$ is small we say that the problem is well conditioned, otherwise we say it's ill conditioned.\\ \ARROW The {\it condition} relation is defined as: \begin{align*} cond(A) = \Vert A \Vert \cdot \Vert A^{-1} \Vert \end{align*} \ARROW Now there are many definitions of different norms... The most popular one (so-called ''column norm''): \begin{align*} \Vert A \Vert_1 = \max_{1 \leq j \leq n} \sum_{i=1}^n \vert a_{i,j} \vert, \end{align*} where $n$ -is the dimension of $A$, $i,j$ are columns and rows numbers. \end{small} \end{frame} \begin{frame}\frametitle{More norms} \begin{small} \ARROW A different norm is a spectral norm: \begin{align*} \Vert A \Vert_2 &= \sqrt{\rho(A^T A)}\\ \rho(M) &= \max \lbrace \vert\lambda_i \vert: \det{M- \lambda I} =0,~i=1,...n \rbrace \end{align*} where $\rho(M)$ - spectral radius of $M$ matrix, $I$ unit matrix, $\lambda_i$ eigenvalues of $M$.\\ \ARROW Row norm: \begin{align*} \Vert A \Vert_{\infty} &= \max_{1 \leq i \leq n} \sum_{j=1}^n \vert a_{i,j} \vert, \end{align*} \begin{exampleblock}{Digression:} \ARROWR Calculation of the matrix norms are not a simple process at all. There are certain class of matrices that make the calculations easier.\\ \ARROWR The spectral norm can be also defined: \begin{align*} cond_2(A) = \frac{\max_{1\leq i \leq n}\vert \lambda_i \vert }{\min_{1\leq i \leq n}\vert \lambda_i \vert }, \end{align*} \end{exampleblock} \end{small} \end{frame} \begin{frame}\frametitle{Example, ill-conditioned matrix} \begin{small} \ARROW The text-book example of wrongly conditioned matrix is the Hilbert matrix: \begin{align*} h_{i,j} = \frac{1}{i+j-1} \end{align*} \ARROWR Example: \begin{align*} h_{i,j}^{4 \times 4} = \begin{pmatrix} 1 & \frac{1}{2} & \frac{1}{3} & \frac{1}{4} \\ \frac{1}{2} & \frac{1}{3} & \frac{1}{4} & \frac{1}{5} \\ \frac{1}{3} & \frac{1}{4} & \frac{1}{5} & \frac{1}{6}\\ \frac{1}{4} & \frac{1}{5} & \frac{1}{6} & \frac{1}{7} \end{pmatrix} \end{align*} \ARROW The condition of this matrix: \begin{align*} cond(A)=\mathcal{O}\left(\frac{e^{3.52N}}{\sqrt{N}}\right) \end{align*} \ARROW For $8 \times 8$ matrix we get: \begin{align*} cond_1(A)=3.387\cdot 10^{10},~~~cond_2(A)=1.526\cdot 10^{10},~~~~~cond_{\infty}(A)=3.387\cdot 10^{10} \end{align*} \ARROW Clearly large numbers ;) \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: Cramer method} \begin{small} \ARROW If $\det{A} \neq 0$ then the solutions are given by: \begin{align*} x_i =\frac{\det{A_i}}{\det{A}} \end{align*} \ARROW So calculate the solutions one needs to calculate $n+1$ determinants. To calculate each determinate one needs $(n-1)n!$ multiplications. \\ \ARROW Putting it all together one needs $(n+1)(n-1)n! = n^{n+2}$ \\ \ARROW Brute force but works ;) \end{small} \end{frame} \begin{frame}\frametitle{Example, ill-conditioned matrix} \begin{small} \ARROW The text-book example of wrongly conditioned matrix is the Hilbert matrix: \begin{align*} h_{i,j} = \frac{1}{i+j-1} \end{align*} \ARROWR Example: \begin{align*} h_{i,j}^{4 \times 4} = \begin{pmatrix} 1 & \frac{1}{2} & \frac{1}{3} & \frac{1}{4} \\ \frac{1}{2} & \frac{1}{3} & \frac{1}{4} & \frac{1}{5} \\ \frac{1}{3} & \frac{1}{4} & \frac{1}{5} & \frac{1}{6}\\ \frac{1}{4} & \frac{1}{5} & \frac{1}{6} & \frac{1}{7} \end{pmatrix} \end{align*} \ARROW The condition of this matrix: \begin{align*} cond(A)=\mathcal{O}\left(\frac{e^{3.52N}}{\sqrt{N}}\right) \end{align*} \ARROW For $8 \times 8$ matrix we get: \begin{align*} cond_1(A)=3.387\cdot 10^{10},~~~cond_2(A)=1.526\cdot 10^{10},~~~~~cond_{\infty}(A)=3.387\cdot 10^{10} \end{align*} \ARROW Clearly large numbers ;) \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: Cramer method} \begin{small} \ARROW If $\det{A} \neq 0$ then the solutions are given by: \begin{align*} x_i =\frac{\det{A_i}}{\det{A}} \end{align*} \ARROW So calculate the solutions one needs to calculate $n+1$ determinants. To calculate each determinate one needs $(n-1)n!$ multiplications. \\ \ARROW Putting it all together one needs $(n+1)(n-1)n! = n^{n+2}$ \\ \ARROW Brutal force but works ;) \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: Gauss method} \begin{small} \ARROW The idea besides the Gauss method is simple: transform the $A x =b$ to get the equvalent matrix $A^{\left[n\right]} x = b^{\left[n\right]}$ where $A^{\left[n\right]}$ is triangular matrix: \begin{align*} A^{\left[n\right]} = \begin{pmatrix} a^{\left[n\right]}_{11} & a^{\left[n\right]_{12}} & ... & a^{\left[n\right]}_{1n}\\ 0 & a^{\left[n\right]}_{22} & ... & a^{\left[n\right]}_{2n} \\ ...\\ 0 & 0 & ... & a^{\left[n\right]}_{nn} \\ \end{pmatrix} \end{align*} \ARROWR The algorithm: \ARROW To do so we calculate the: $d^{\left[1\right]}_{i,1}=\frac{a^{\left[1\right]}_{i1}}{a^{\left[1\right]}_{11}}$ \\ \ARROW The first row multiplied by the $d^{\left[1\right]}_{i,1}$ we subtract from the $i^{th}$ row. \ARROW After this we get: \begin{align*} \begin{pmatrix} a^{\left[ 1 \right]}_{11} & a^{\left[ 1 \right]_{12}} & ... & a^{\left[ 1 \right]}_{1n}\\ 0 & a^{\left[ 1 \right]}_{22} & ... & a^{\left[ 1 \right]}_{2n} \\ ...\\ 0 & a^{\left[ 1 \right]}_{n2} & ... & a^{\left[ 1 \right]}_{nn} \\ \end{pmatrix} \overrightarrow{x}= \begin{pmatrix} b^{\left[ 1 \right]}_{1} \\ b^{\left[ 1 \right]}_{1} \\ ... \\ b^{\left[ 1 \right]}_{1} \\ \end{pmatrix} \end{align*} \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: Gauss method 2} \begin{small} \ARROW Now one needs to repeat the above n times moving each time row down. \begin{alertblock}{Cons:} \ARROW The algoright can be stooped if you divide by zero.\\ \ARROW The method is very efficient to accumulate numerical errors. \end{alertblock} \begin{exampleblock}{Pros:} \ARROWR The number of needed floating point operations is less then Cramer.\\ \ARROWR Example for 15 equations: $1345$ vs $5 \cdot 10^{12}$. \end{exampleblock} \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: modified Gauss method} \begin{small} \ARROW The biggest disadvantage of Gauss method is the fact that we can have zero elements :( \\ \ARROW The modified method fixes this problem :)\\ \ARROW The modification is as follows: \begin{itemize} \item In each step before we do the elimination we look for main element: \begin{align*} \vert a_{mk}^{\vert k \vert } = \max \lbrace \vert a_{jk}^{\vert k \vert} : j=k, k+1,..., n \rbrace \end{align*} \item We exchange the rows $m$ and $k$. \item We do the standard elimination. \end{itemize} \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: Jordan elimination method} \begin{small} \ARROW The Jordan elimination method is similar to Guass method but the idea is to transform the matrix from $Ax =b$ to $I x = b^{\lfloor n+1 \rfloor}$. \begin{enumerate} \item We start from eliminating the as in Gauss method in the first row. \item When we move to the second row we eliminate the $x_2$ element also from the first raw. \item The third raw we eliminate the first and second raw. \item We repeat this $n$ times. \end{enumerate} \ARROW After this we will get new system $I x = b^{\lfloor n+1 \rfloor}$. \\ \ARROW The $b^{\lfloor n+1 \rfloor}$ is already the solution! No need to do more. \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: LU method} \begin{small} \ARROW The most popular of the exact methods is so-called LU method.\\ \ARROW The idea is very simple; we represent the matrix in a form: \begin{align*} A = & \begin{pmatrix} a_{11} & a_{12} & a_{13} & ... & a_{1n} \\ a_{21} & a_{22} & a_{23} & ... & a_{2n} \\ a_{31} & a_{32} & a_{33} & ... & a_{2n} \\ & & ... & & \\ _{n1} & a_{n2} & a_{n3} & ... & a_{nn} \\ \end{pmatrix} = \\ & \begin{pmatrix} 1 & 0 & 0 & ... & 0 \\ l_{21} & 1 & 0 & ... & 0 \\ l_{31} & l_{32} & 1 & ... &0 & \\ & & ... & & \\ l_{n1} & l_{n2} & l_{n3} & ... & l_{nn}\\ \end{pmatrix} \cdot \begin{pmatrix} u_{11} & u{12} & u_{13} & ... & u_{1n} \\ 0 & u{22} & u_{23} & ... & u_{2n} \\ 0 & 0 & u_{33} & ... & u_{3n} \\ & & ... & & \\ 0 & 0 & 0 & ... & u_{nn} \end{pmatrix} = L \cdot U \end{align*} \ARROW After this decomposition we need to solve: \begin{align*} \begin{cases} L y & =b\\ U x & =y \end{cases} \end{align*} \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: LU method, algorithms} \begin{small} \ARROW We solve the following matrix: $A^{\lfloor 1 \rfloor} x = b^{\lfloor1 \rfloor}$\\ \ARROW We start by preparing the matrix $L^{\lfloor 1 \rfloor}$: \begin{align*} L^{\lfloor 1 \rfloor }=\begin{pmatrix} 1 & 0 & 0 & ... & 0 \\ -l_{21} & 1 & 0 & ... & 0\\ -l_{31} & 0 & 1 & ... & 0\\ & & ... & & \\ -l_{n1} & 0 & 0 & ... & 1 \end{pmatrix} \end{align*} where the $l_{i1}$ are defined: \begin{align*} l_{i1} = \frac{a_{i1}^{\lfloor 1 \rfloor}}{a_{11}^{\lfloor 1 \rfloor}} \end{align*} \ARROW Now we take our base equation by $L^{\lfloor 1 \rfloor }$: \begin{align*} L^{\lfloor 1 \rfloor } A^{\lfloor 1 \rfloor } x = L^{\lfloor 1 \rfloor } b^{\lfloor1 \rfloor} \end{align*} \ARROW We get the a new system: \begin{align*} A^{\lfloor 2 \rfloor} x = b^{\lfloor2 \rfloor}~~~~~~~~\Leftrightarrow~~~~~~~~~~L^{\lfloor 1 \rfloor } A^{\lfloor 1 \rfloor } x = L^{\lfloor 1 \rfloor } b^{\lfloor1 \rfloor} \end{align*} \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: LU method, algorithms} \begin{small} \ARROW In the second step we construct the $L^{\lfloor 2 \rfloor}$ in the form: \begin{align*} L^{\lfloor 2 \rfloor}=\begin{pmatrix} 1 & 0 & 0 & ... & 0 \\ 0 & 1 & 0 & ... & 0\\ 0 & -l_{32} & 1 & ... & 0\\ & & ... & & \\ 0 & -l_{n2} & 0 & ... & 1 \end{pmatrix} \end{align*} where : \begin{align*} l_{i2} = \frac{a_{i2}^{\lfloor 2 \rfloor}}{a_{22}^{\lfloor 2 \rfloor}} \end{align*} \ARROW Now we take the $A^{\lfloor 2 \rfloor} x = b^{\lfloor2 \rfloor}$ and multiply it by $L^{\lfloor 2 \rfloor}$: \begin{align*} L^{\lfloor 2 \rfloor } A^{\lfloor 2 \rfloor } x = L^{\lfloor 2 \rfloor } b^{\lfloor2 \rfloor} \end{align*} \ARROW We get the a new system: \begin{align*} A^{\lfloor 3 \rfloor} x = b^{\lfloor3 \rfloor}~~~~~~~~\Leftrightarrow~~~~~~~~~L^{\lfloor 2 \rfloor } L^{\lfloor 1 \rfloor } A^{\lfloor 1 \rfloor } x = L^{\lfloor 2 \rfloor } L^{\lfloor 1 \rfloor } b^{\lfloor1 \rfloor} \end{align*} \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: LU method, algorithms} \begin{small} \ARROW Now the we repeat the above steps $n-1$ times after which we get: \begin{align*} & L^{\lfloor n-1 \rfloor } L^{\lfloor n-2 \rfloor } ... L^{\lfloor 2 \rfloor } L^{\lfloor 1 \rfloor } A^{\lfloor 1 \rfloor } = A^{\lfloor n \rfloor } = U\\ & L^{\lfloor n-1 \rfloor } L^{\lfloor n-2 \rfloor } ... L^{\lfloor 2 \rfloor } L^{\lfloor 1 \rfloor } b^{\lfloor1 \rfloor} = b^{\lfloor n \rfloor} \end{align*} \ARROW From the above we can calculate: \begin{align*} A^{\lfloor 1 \rfloor } = \left( L^{\lfloor 1 \rfloor } \right)^{-1} \left( L^{\lfloor 2 \rfloor } \right)^{-1} ... \left( L^{\lfloor n-1 \rfloor } \right)^{-1} A^{\lfloor n \rfloor } \end{align*} \ARROW So the matrix $L$ we search for is: \begin{align*} L= \left( L^{\lfloor 1 \rfloor } \right)^{-1} \left( L^{\lfloor 2 \rfloor } \right)^{-1} ... \left( L^{\lfloor n-1 \rfloor } \right)^{-1} \end{align*} \ARROW The $\left( L^{\lfloor k \rfloor } \right)^{-1}$ can be easy calculated: \begin{footnotesize} \begin{align*} \left( L^{\lfloor k \rfloor } \right)^{-1} = \begin{pmatrix} 1 & 0 & ... & 0 & ... & 0 \\ 0 & 1 & ... & 0 & ... & 0 \\ & & & ... & & \\ 0 & 0 & ... & 1 & ... & 0 \\ 0 & 0 & ... & l_{k+1 k} & ... & 0 \\ 0 & 0 & ... & l_{k+2 k} & ... & 0 \\ & & & ... & & \\ 0 & 0 & ... & l_{n k} & ... & 1 \\ \end{pmatrix} \end{align*} \end{footnotesize} \end{small} \end{frame} \begin{frame}\frametitle{Exact methods: LU method, algorithms} \begin{small} \ARROW Now the only thing left is to solve the simple linear system: \begin{align*} \begin{cases} L y & =b\\ U x & =y \end{cases} \end{align*} \ARROW Because of the triangular matrix the solution is straightforward: \begin{align*} y_1=\frac{b_1}{L_11},~~~~~~~~~y_i=\frac{b_1 - \sum_{j=1^{i-1}} L_{ij} y_j }{L_{ii}}~~{ i\geq 2} \end{align*} \end{small} \end{frame} \begin{frame}\frametitle{Summary} \begin{small} \ARROW This lecture we learn the exact methods of solving linear equation system.\\ \ARROW The three most popular one are: Gauss, Jordan, LU.\\ \ARROW By default the LU method should be used.\\ \ARROW And remember: be sure the system is well conditioned! \end{small} \end{frame} \backupbegin \begin{frame}\frametitle{Backup} \end{frame} \backupend \end{document}