Contents

MIT 18.06(Linear Algebra) L2 Note

1. Mind Map

https://raw.githubusercontent.com/loss4wang/wx_imagehost/main/LA_Lecture_2.png

2. Reading Notes(2.2-2.3)

2.2 The Idea of Elimination

A systematic way to solve linear equations: elimination

  • GOAL: produce an upper triangular system
  • The system solved from the bottom upwards: back substitution
  • Pivot : first nonzero in the row that does the elimination
  • Multiplier : (entry to eliminate) divided by (pivot)

Breakdown of Elimination

Failure: The method might ask us to divide by zero.

  • Permanent failure with no solution.
  • Failure with infinitely many solutions.

  • Elimination leads to an equation $0 \neq 0$ (no solution) or 0 = 0 (many solutions)

Success comes with n pivots. But we may have to exchange the n equations.

Three Equations in Three Unknowns

Goal: Forward elimination is complete from A to U.

Elimination from A to U

  • Column 1. Use the first equation to create zeros below the first pivot.
  • Column2. Use the new equation2 to create zeros below the second pivot.
  • Columns 3 to n. Keep going to find all n pivots and the upper triangular U.

2.3 Elimination Using Matrices

Matrices times Vectors and Ax = b

  • $Ax=b$
    • Ax is a combination of columns of A
    • Components of Ax are dot products with rows of A.

The Matrix Form of One Elimination Step

$Ax=b: \begin{bmatrix}2&4&-2 \cr 4&9&-3 \cr -2&-3&7 \end{bmatrix} \begin{bmatrix} -1 \cr 2 \cr 2 \end{bmatrix} =\begin{bmatrix}2 \cr 8 \cr 10 \end{bmatrix}$

  • First step: Subtract 2*Row1 from Row2
    • Elimination matrix is $E = \begin{bmatrix}1&0&0 \cr -2&1&0 \cr 0&0&1 \end{bmatrix}$
    • $b_{new} = Eb$
    • $\begin{bmatrix}1&0&0 \cr -2&1&0 \cr 0&0&1 \end{bmatrix} \begin{bmatrix} 2 \cr 8 \cr 10 \end{bmatrix} =\begin{bmatrix}2 \cr 4 \cr 10 \end{bmatrix}$
    • $\begin{bmatrix}1&0&0 \cr -2&1&0 \cr 0&0&1 \end{bmatrix} \begin{bmatrix}b_1 \cr b_2 \cr b_3 \end{bmatrix} = \begin{bmatrix}b_1 \cr -2b_1+b_2 \cr b_3 \end{bmatrix}$
  • I: identity matrix $\begin{bmatrix}1&0&0 \cr 0&1&0 \cr 0&0&1 \end{bmatrix}$
  • E :elementary matrix or elimination matrix
    • $E_{ij}$ has extra nonzero entry $-l$ in the i,j position. Then $E_{ij}$ subtracts a multiple $l$ of row j from row i.
    • $E_{31} = \begin{bmatrix}1&0&0 \cr 0&1&0 \cr -l&0&1 \end{bmatrix}$
    • The purpose of $E_{31}$ **is to produce a zero in the ( 3, 1) position of the matrix.**
    • Products and inverses are especially clear for E’s. It is those two ideas that the book will use.

Matrix Multiplication

Q: How do we multiply two matrices?

  • $E(Ax) =Eb \quad{also}\quad (EA)x = Eb$
  • Associative law is true
  • Commutative law is false
  • E on the rights acts on the columns of A
  • E on the left acts on the rows of A

Matrix multiplication:

$AB = A [b_1,b_2,b_3]=[Ab_1,Ab_2,Ab_3]$

The Matrix $P_{ij}$ for a Row Exchange

Permutation Matrix

  • A row exchange is needed when zero is in the pivot position.

$$ \begin{bmatrix}1&0&0 \cr 0&0&1 \cr 0&1&0 \end{bmatrix} \begin{bmatrix}2&4&1 \cr 0&0&3 \cr 0&6&5 \end{bmatrix} = \begin{bmatrix}2&4&1 \cr 0&6&5 \cr 0&0&3 \end{bmatrix} $$

  • Row Exchange Matrix $P_{ij}$ is the identity matrix with rows i and j reversed.

The Augmented Matrix

$$ \text{Augmented matrix} [A\quad b] = \begin{bmatrix} 2&4&-1&2 \cr 4&9&-3&4 \cr -2 &-3&7&10 \end{bmatrix} $$