2.1 Linear Equations Picture

Row Picture

2 by 2 equations

Two equations, Two unknowns

\[\begin{matrix} x - 2y = 1 \\ 3x + 2y = 11 \end{matrix}
\]

The row picture shows two lines meeting at a single point(the solution).

3 by 3 equations

Three equations, Three unknowns

\[Ax=b \ \ => \ \ \begin{matrix} x + 2y + 3z = 6 \\ 2x + 5y + 2z = 4 \\ 6x - 3y + z = 2\end{matrix}
\]

The row picture shows three planes meeting at a single point.

Column Picture

2 by 2 equations

Two equations, Two unknowns

\[\begin{matrix} x - 2y = 1 \\ 3x + 2y = 11 \end{matrix} =>x\left[ \begin{matrix} 1\\ 3 \\ \end{matrix} \right] + y\left[ \begin{matrix} -2\\ 2 \\ \end{matrix} \right] =\left[\begin{matrix} 1 \\ 11 \end{matrix}\right] = b
\]

The column picture combines the column vectors on the left side to produce the vector b on the right side.

(The left side of the vector equation is a linear combination of the columns)

3 by 3 equations

Three equations, Three unknowns

\[\begin{matrix} x + 2y + 3z = 6 \\ 2x + 5y + 2z = 4 \\ 6x - 3y + z = 2 \end{matrix} =>x\left[ \begin{matrix} 1\\ 2\\6 \\ \end{matrix} \right] + y\left[ \begin{matrix} 2\\ 5 \\-3 \\ \end{matrix} \right] + z\left[ \begin{matrix} 3\\ 2 \\1 \\ \end{matrix} \right] =\left[\begin{matrix} 4 \\ 6 \\ 2\end{matrix}\right] = b
\]

The column picture combines three columns to produce b,the coefficients (x,y,z) = (0,0,2).

2.2 Elimination

2.2.1 Gaussian Elimination

  • Column 1 : Use the first equation to create zeros below the first pivot.
  • Column 2 : Use the new equation 2 to create zeros below the second pivot.
  • Column 3 to n : Keep going to find all n pivots and the upper triangular U.
2 by 2

Multiply equation 1 by 3, and Subtract from equation 2.

\[(Before) \ \ \begin{matrix} x - 2y = 1 \\ 3x + 2y = 11 \end{matrix} ==> (After) \ \ \begin{matrix} x - 2y = 1 \\ 8y = 8 \end{matrix}
\]

3 by 3
\[Ax = b \ ==> \ \begin{matrix} 2x + 4y - 2z = 2 \\ 4x + 9y - 3z = 8 \\ -2x-3y+7z=10 \end{matrix}
\]

Elimination Steps

step1 : Subtract 2 times equation 1 from equation 2.

\[\begin{matrix} 2x + 4y - 2z = 2 \\
\quad \quad \quad y + z = 4 \\
-2x-3y+7z=10 \end{matrix}
\]

step2 : Subtract -1 times equation 1 from equation 3.

\[\begin{matrix} 2x + 4y - 2z = 2 \\
\quad \quad \quad y + z = 4 \\
\quad \quad \quad \quad y+5z=12 \end{matrix}
\]

step3 : Subtract new equation 2 from new equation 3.

\[\quad \begin{matrix} 2x + 4y - 2z = 2 \\
\quad \quad \quad \quad y + z = 4 \\
\quad \quad \quad \quad \quad 4z = 8 \end{matrix}
==>Ux = c
\]

U is upper triangular.

Back substitution

z = 2 --> y = 2 --> x = -1

2.2.2 Elimination-Matrix

Elimination multiplies Ax=b by \(E_{21} , E_{31} , E_{41}, ..., E_{n1}\), then \(E_{32} , E_{42}, ..., E_{n2}\) and onward.

  • \(E =E_{21} ,..., E_{n1},..., E_{n2},...,E_{n(n-1)}\) , \(EA = [Ea_1...Ea_n]\)
  • Augmented matrix : \(E[A\ \ b] = [EA\ \ Eb]\)

example:

\[Ax = b \\
\Downarrow \\
\begin{matrix} 2x_1 + 4x_2 - 2x_3 = 2 \\ 4x_1 + 9x_2 - 3x_3 = 8 \\ -2x_1-3x_2+7x_3=10 \end{matrix} \\
\Downarrow \\
\left[ \begin{matrix} 2&4&-2 \\ 4&9&-3 \\ -2&-3&7 \end{matrix} \right]
\left[ \begin{matrix} x_1\\x_2\\x_3 \end{matrix} \right] =
\left[ \begin{matrix} 2\\8\\10 \end{matrix} \right] \\
\Downarrow \\
\left[ \begin{matrix} 1&0&0 \\ -2&1&0 \\ 0&0&1 \end{matrix} \right]
\left[ \begin{matrix} 2&4&-2 \\ 4&9&-3 \\ -2&-3&7 \end{matrix} \right]
\left[ \begin{matrix} x_1\\x_2\\x_3 \end{matrix} \right] =
\left[ \begin{matrix} 1&0&0 \\ -2&1&0 \\ 0&0&1 \end{matrix} \right]
\left[ \begin{matrix} 2\\8\\10 \end{matrix} \right] \\
\Downarrow \\
\left[ \begin{matrix} 2&4&-2 \\ 0&1&1 \\ -2&-3&7 \end{matrix} \right]
\left[ \begin{matrix} x_1\\x_2\\x_3 \end{matrix} \right] =
\left[ \begin{matrix} 2\\4\\10 \end{matrix} \right] \\
\Downarrow \\
\left[ \begin{matrix} 1&0&0 \\ 0&1&0 \\ 1&0&1 \end{matrix} \right]
\left[ \begin{matrix} 2&4&-2 \\ 0&1&1 \\ -2&-3&7 \end{matrix} \right]
\left[ \begin{matrix} x_1\\x_2\\x_3 \end{matrix} \right] =
\left[ \begin{matrix} 1&0&0 \\ 0&1&0 \\ 1&0&1 \end{matrix} \right]
\left[ \begin{matrix} 2\\4\\10 \end{matrix} \right] \\
\Downarrow \\
\left[ \begin{matrix} 2&4&-2 \\ 0&1&1 \\ 0&1&5 \end{matrix} \right]
\left[ \begin{matrix} x_1\\x_2\\x_3 \end{matrix} \right] =
\left[ \begin{matrix} 2\\4\\12 \end{matrix} \right] \\
\Downarrow \\
\left[ \begin{matrix} 1&0&0 \\ 0&1&0 \\ 0&-1&1 \end{matrix} \right]
\left[ \begin{matrix} 2&4&-2 \\ 0&1&1 \\ 0&1&5 \end{matrix} \right]
\left[ \begin{matrix} x_1\\x_2\\x_3 \end{matrix} \right] =
\left[ \begin{matrix} 1&0&0 \\ 0&1&0 \\ 0&-1&1 \end{matrix} \right]
\left[ \begin{matrix} 2\\4\\12 \end{matrix} \right] \\
\Downarrow \\
\left[ \begin{matrix} 2&4&-2 \\ 0&1&1 \\ 0&0&4 \end{matrix} \right]
\left[ \begin{matrix} x_1\\x_2\\x_3 \end{matrix} \right] =
\left[ \begin{matrix} 2\\4\\8 \end{matrix} \right] \\
\Downarrow Back \ \ substitution \\
x_3 = 2 , x_2 = 2, x_1 = -1
\]

2.3 Rules for Matrix Operations

2.3.1 Matrix Multiplication

Matrices A with n columns multiply matrices B with n rows : \(A_{m \times n} B_{n \times p} = C_{m \times p}\)

The regular way

The entry in row i and column j of AB is (row i of A) \(\cdot\) (column j of B): \((AB)_{ij}=a_{i1}b_{1j} + a_{i2}b_{2j}+...+a_{in}b_{nj}\)

\[\left[ \begin{matrix} * \\ a_{i1}&a_{i2}&...&a_{in} \\ * \\* \end{matrix} \right]
\left[ \begin{matrix} *&b_{1j}&*&*\\ &b_{2j}&&\ \\ &\vdots&& \\ &b_{nj}&& \end{matrix} \right]=
\left[ \begin{matrix} &&*&& \\ *&*&(AB)_{ij}&*&* \\ &&*&& \\&&*&& \end{matrix} \right]
\]
The column way

Each column of AB is a combination of the columns of A.

Matrix A times every column of B : \(A[b_1...b_p]=[Ab_1...Ab_p]\)

The row way

Every row of AB is a combination of the rows of B

Every row of A times matrix B : \(\left[\begin{matrix} a_1 \\ a_2 \\ \vdots \\a_n \end{matrix}\right]B=\left[\begin{matrix} a_1B \\ a_2B \\ \vdots \\a_nB \end{matrix}\right]\)

The columns multiply rows

Multiply columns 1 to n of A times rows 1 to n of B. Add those matrices.

\[\left[\begin{matrix} col_1&\cdots&col_n \\ \vdots&\vdots&\vdots \end{matrix}\right]
\left[\begin{matrix} row_1&\cdots \\ \vdots&\vdots \\row_n&\cdots \end{matrix}\right]
=(col_1)(row_1)+...+(col_n)(row_n)
\]
Block Multiplication

A and B cut into blocks(which are small matrices).

\[A = \left[\begin{matrix} A_1&A_2\\ A_3&A_4 \end{matrix}\right] \\
B = \left[\begin{matrix} B_1&B_2\\ B_3&B_4 \end{matrix}\right] \\
AB =\left[\begin{matrix} A_1&A_2\\ A_3&A_4 \end{matrix}\right]
\left[\begin{matrix} B_1&B_2\\ B_3&B_4 \end{matrix}\right] =
\left[\begin{matrix} A_1B_1 + A_2B_3&A_1B_2 + A_2B_4\\ A_3B_1 + A_4B_3&A_2B_2 + A_4B_4\end{matrix}\right]
\]

2.3.2 The Laws for Matrix Operations

Additions

Commutative law : A + B = B + A

Distributive law : c(A + B) = cA + cB

Associative law : A + (B + C) = (A + B) + C

Multiply

Commutative law is usually broken : \(AB \neq BA\)

Distributive law : (A + B)C = AC + BC or C(A + B) = CA + CB

Associative law : A (B C) = (A B) C

2.4 Inverse Matrices

The matrix A is invertible if there exists a matrix \(A^{-1}\) that "inverts" A :

\[A^{-1}A = I \quad and \quad AA^{-1}=I
\]
  • A is invertible if and only if it has n pivots (row exchanges allowed).
  • If Ax = 0 for a nonzero vector x, then A has no inverse.
  • The inverse of AB is the reverse product \(B^{-1}A^{-1}\),and \((ABC)^{-1}=C^{-1}B^{-1}A^{-1}\).
  • Diagonally dominant matrices are invertible.Each \(|a_{ii}|\)dominates its row.

Gauss-Jordan Method

\[[A \quad I] \quad reduce \quad to \quad [I \quad A^{-1}]
\]

example $A = \left[ \begin{matrix} 2&3 \ 4&7 \end{matrix}\right] $:

\[[A \quad I] = \left[ \begin{matrix} 2&3&1&0 \\ 4&7&0&1 \end{matrix}\right] \\
\Downarrow \\
[U \quad L^{-1}]=\left[ \begin{matrix} 2&3&1&0 \\ 0&1&-2&1 \end{matrix}\right] \quad \\
\Downarrow \\
\left[ \begin{matrix} 2&0&7&-3 \\ 0&1&-2&1 \end{matrix}\right] \\
\Downarrow \\
[I \quad A^{-1}]=\left[ \begin{matrix} 1&0&7/2&-3/2 \\ 0&1&-2&1 \end{matrix}\right] \quad \\
\]

2.5 Factorization : A = LU

Gaussian elimination (with no row exchanges) factors A into L times U,the factors L and U are triangular matrices, and L include all their inverse.

\[A = LU \quad (which \quad L \rightarrow lower \quad triangular, \quad U \rightarrow upper \quad trangular)
\]
\[(E_{n(n-1)}...E_{31}E_{21})A = U \\
\Downarrow \\
(E_{21}^{-1}E_{31}^{-1}...E_{n(n-1)}^{-1})(E_{n(n-1)}...E_{31}E_{21})A = (E_{21}^{-1}E_{31}^{-1}...E_{n(n-1)}^{-1})U \\
\Downarrow \\
A = LU \\
\]

example \(A = \left[ \begin{matrix} 2&1&0 \\ 1&2&1 \\ 0&1&2 \end{matrix}\right] =
\left[ \begin{matrix} 1&0&0 \\ 1/2&1&0 \\ 0&2/3&1 \end{matrix}\right]
\left[ \begin{matrix} 2&1&0 \\ 0&3/2&1 \\ 0&0&4/3 \end{matrix}\right] = LU\)

The triangular factorization can be written : \(A = LU \rightarrow A=LDU\), that D is a diagonal matrix contains the pivots.

Split U into \(DU=\left[ \begin{matrix} d_1&&& \\ &d_2&& \\ &&\ddots \\ &&&d_n \end{matrix}\right]\left[ \begin{matrix} 1&u_{12}/d_1&u_{13}/d_1&\cdots \\ &1&u_{23}/d_2&\vdots \\ &&\ddots \\ &&&1 \end{matrix}\right]\)

example:

\[A = \left[ \begin{matrix} 2&1&0 \\ 1&2&1 \\ 0&1&2 \end{matrix}\right] \\ =
\left[ \begin{matrix} 1&0&0 \\ 1/2&1&0 \\ 0&2/3&1 \end{matrix}\right]
\left[ \begin{matrix} 2&1&0 \\ 0&3/2&1 \\ 0&0&4/3 \end{matrix}\right] \\ =
\left[ \begin{matrix} 1&0&0 \\ 1/2&1&0 \\ 0&2/3&1 \end{matrix}\right]
\left[ \begin{matrix} 2&0&0 \\ 0&3/2&0 \\ 0&0&4/3 \end{matrix}\right]\left[ \begin{matrix} 1&1/2&0 \\ 0&1&2/3 \\ 0&0&1 \end{matrix}\right]= LDU
\]

Keys

  • The lower triangular L contains the number \(l_{ij}\) that multiply pivot rows, going from A to U. The product LU adds those rows back to recover A.
  • On the right side we solve Lc = b (forward) and Ux=c (backward).
  • Cost : the left side costs \(1/3(n^3 -n)\) multiplications and subtractions,the right side costs \(n^2\) multiplications and subtractions.

2.6 Transposes and Permutations

Transposes

The columns of \(A^{T}\) are the rows of A

\[(A^{T})_{ij} = A_{ji}
\]

If \(A = \left [ \begin{matrix} 1&2&3 \\ 0&0&4 \end{matrix}\right]\) then \(A^{T} = \left [ \begin{matrix} 1&0 \\ 2&0 \\ 3&4 \end{matrix}\right]\)

Sum : \((A+B)^{T} = A^{T} + B^{T}\)

Product : \((AB)^{T} = B^{T}A^{T}\)

Inverse : \((A^{T})^{-1} = (A^{-1})^{T}\)

Symmetric matrix (\(S^T=S\)):\(U = L^T \rightarrow S = LDU = LDL^T\)

Permutations

A permutation matrix P has the rows of the identity I in any order, \(P_{ij}\) is constructed by exchanging two row i and j of \(I\),and there are \(n!\) permuataion matrices of order n.

3 by 3 permuation matrices:

\[I = \left [ \begin{matrix} 1&& \\ &1& \\ &&1 \end{matrix}\right] \quad
P_{21} = \left [ \begin{matrix} &1& \\ 1&& \\ &&1 \end{matrix}\right] \quad
P_{31} = \left [ \begin{matrix} &&1\\ &1& \\ 1&& \end{matrix}\right] \\
P_{32} = \left [ \begin{matrix} 1&&\\ &&1 \\ &1& \end{matrix}\right] \quad
P_{32}P_{21} = \left [ \begin{matrix} &1&\\ &&1 \\ 1&& \end{matrix}\right] \quad
P_{21}P_{32} = \left [ \begin{matrix} &&1\\ 1&& \\ &1& \end{matrix}\right]
\]
  • If A is invertible then a permutation P will reorder its rows for PA=LU.
  • A permutation matrix P has a 1 in each row and column, and \(P^T = P^{-1}\).

2. Solving Linear Equations的更多相关文章

  1. Linear Equations

    4.1 Linear Equations with One Independent Variable

  2. Linear Equations in Linear Algebra

    Linear System Vector Equations The Matrix Equation Solution Sets of Linear Systems Linear Indenpende ...

  3. 线性代数导论 | Linear Algebra 课程

    搞统计的线性代数和概率论必须精通,最好要能锻炼出直觉,再学机器学习才会事半功倍. 线性代数只推荐Prof. Gilbert Strang的MIT课程,有视频,有教材,有习题,有考试,一套学下来基本就入 ...

  4. Java基础常见英语词汇

    Java基础常见英语词汇(共70个) ['ɔbdʒekt] ['ɔ:rientid]导向的                             ['prəʊɡræmɪŋ]编程 OO: object ...

  5. 看到了必须要Mark啊,最全的编程中英文词汇对照汇总(里面有好几个版本的,每个版本从a到d的顺序排列)

    java:  第一章: JDK(Java Development Kit) java开发工具包 JVM(Java Virtual Machine) java虚拟机 Javac  编译命令 java   ...

  6. (转)Awesome Courses

    Awesome Courses  Introduction There is a lot of hidden treasure lying within university pages scatte ...

  7. 专业英语词汇(Java)

    abstract (关键字)             抽象 ['.bstr.kt] access                            vt.访问,存取 ['.kses]‘(n.入口, ...

  8. Lua的各种资源1

    Libraries And Bindings     LuaDirectory > LuaAddons > LibrariesAndBindings This is a list of l ...

  9. JAVA常用单词

    柠檬学院Java 基础常见英语词汇(共 70 个)OO: object-oriented ,面向对象 OOP: object-oriented programming,面向对象编程JDK:Java d ...

  10. java常用英语单词

    abstract (关键字) 抽象 ['.bstr.kt] access vt.访问,存取 ['.kses]'(n.入口,使用权) algorithm n.算法 ['.lg.riem] annotat ...

随机推荐

  1. EL表达式处理字符串

    示例 JSP页面页头添加 <%@ taglib uri="/WEB-INF/taglib/c.tld" prefix="c"%> <%@ ta ...

  2. 【Azure 应用服务】使用App Service for Linux/Container时,如果代码或Container启动耗时大于了230秒,默认会启动失败。

    问题描述 使用App Service for Linux/Container时,从Docker的日志中,我们可以看见有 warmup 行为,而此行为默认时间为230秒,如果超出了这个时间,就会导致Co ...

  3. Java 线程通信 例子:使用俩个线程打印1-100.线程1 线程2 交替打印

    1 package bytezero.threadcommunication; 2 3 /** 4 * 线程通信的例子:使用俩个线程打印1-100.线程1 线程2 交替打印 5 * 6 * 涉及到的三 ...

  4. Java 对属性赋值的位置 执行的先后顺序

    1 package com.bytezreo.block; 2 3 /** 4 * 5 * @Description 对属性赋值的位置: 6 * @author Bytezero·zhenglei! ...

  5. C++ //常用查找算法 find_if

    1 //常用查找算法 find_if 2 #include<iostream> 3 #include<string> 4 #include<vector> 5 #i ...

  6. OpenCV开发笔记(七十六):相机标定(一):识别棋盘并绘制角点

    前言   知道图像畸变矫映射的原理之后,那么如何得到相机的内参是矫正的第一步,内参决定了内参矩阵(中心点.焦距等),用内参矩阵才能计算出投影矩阵,从而将原本畸变的图像矫正为平面投影图像.  本篇描述了 ...

  7. MinGW编译Python至pyd踩坑整理

    不需要安装VS工具,pyd使用说明. 用scoop自动安装配置MinGw 需要魔法,用包管理scoop安装不需要手动配置.这一步可以自行下载mingw64然后手动配置. scoop install m ...

  8. java多线程的锁整理

    参考,欢迎点击原文:https://www.jianshu.com/p/473a3d96a1b0(java锁总结) https://www.jianshu.com/p/dcabdf695557(Ree ...

  9. 首届实时渲染3D动画创作大赛结果公布,看大奖英伟达A6000花落谁家!

    根据评选标准,经过评委组层层选拔,首届实时渲染3D动画创作大赛,最终获奖结果出炉啦. 本次赛事报名人数达212人,入围作品共40份,其中Omniverse组11份,专业组15份,学生组14份.在宣布最 ...

  10. http内网穿透CYarp[开源]

    0 前言 在物联网领域中,mqtt消息一直是海量设备连接到平台的标配协议,而平台向移动端开放的操作接口往往是http协议,这就要求平台为两种协议作消息一一适配.在某些情况下,这些设备是有操作系统的li ...