353 lines
10 KiB
TeX
353 lines
10 KiB
TeX
\subsection{Recap}
|
|
Given the linear equation $a_1x_1 + a_2x_2 + ... + a_nx_n = b$
|
|
|
|
\begin{enumerate}
|
|
\item $a_1 = a_2 = ... = a_n = b = 0$ zero equation
|
|
|
|
Solution: $x_1 = t_1, x_2 = t_2, ... = x_n = t_n$
|
|
|
|
\item $a_1 = a_2 = ... = a_n = 0 \neq b$ inconsistent
|
|
|
|
No Solution
|
|
|
|
\item Not all $a_1 ... a_n$ are zero.
|
|
|
|
Set $n-1$ of $x_i$ as params, solve for last variable
|
|
\end{enumerate}
|
|
|
|
\subsection{Elementary Row Operations Example}
|
|
|
|
\begin{center}
|
|
|
|
\begin{minipage}{.3\linewidth}
|
|
\systeme{x+y+3z = 0, 2x-2y+2z=4, 3x+9y=3}
|
|
\end{minipage}%
|
|
\begin{minipage}{.3\linewidth}
|
|
\begin{equation*}
|
|
\begin{amatrix}{3}
|
|
1 & 1 & 3 & 0 \\
|
|
2 & 2 & 2 & 4 \\
|
|
3 & 9 & 0 & 3
|
|
\end{amatrix}
|
|
\end{equation*}
|
|
\end{minipage}
|
|
\end{center}
|
|
|
|
\subsection{Row Equivalent Matrices}
|
|
2 Augmented Matrices are row equivalent if one can be obtained from the other by a series of elementary row operations
|
|
|
|
Given a augmented matrix $A$, how to find a row equivalent augmented matrix B of which is of a \textbf{simple} form?
|
|
|
|
\subsection{Row Echelon Form}
|
|
|
|
\begin{defn}[Row Echelon Form (Simple)]
|
|
Augmented Matrix is in row-echelon form if
|
|
\begin{itemize}
|
|
\item Zero rows are grouped together at the bottom
|
|
\item For any 2 successive nonzero rows, The first nonzero number in the lower row appears to the right of the first nonzero number on the higher row
|
|
$\begin{amatrix}{4}
|
|
0 & 0 & 1 & 2 & 3 \\
|
|
0 & 0 & 0 & 1 & 2 \\
|
|
\end{amatrix}$
|
|
\item Leading entry if a nonzero row is a \textbf{pivot point}
|
|
\item Column of augmented matrix is called
|
|
\begin{itemize}
|
|
\item \textbf{Pivot Column} if it contains a pivot point
|
|
\item \textbf{Non Pivot Column} if it contains no pivot point
|
|
\end{itemize}
|
|
\item Pivot Column contains exactly 1 pivot point
|
|
|
|
\# of pivots = \# of leading entries = \# of nonzero rows
|
|
\end{itemize}
|
|
\end{defn}
|
|
|
|
Examples of row echlon form:
|
|
|
|
\begin{equation*}
|
|
\begin{amatrix}{2}
|
|
3 & 2 & 1 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{2}
|
|
1 & -1 & 0 \\
|
|
0 & 1 & 0 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{2}
|
|
2 & 1 & 0 \\
|
|
0 & 1 & 0 \\
|
|
0 & 0 & 1 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{3}
|
|
1 & 2 & 3 & 4 \\
|
|
0 & 1 & 1 & 2 \\
|
|
0 & 0 & 2 & 3 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{4}
|
|
0 & 1 & 2 & 8 & 1 \\
|
|
0 & 0 & 0 & 0 & 3 \\
|
|
0 & 0 & 0 & 0 & 0 \\
|
|
0 & 0 & 0 & 0 & 0 \\
|
|
\end{amatrix}
|
|
\end{equation*}
|
|
|
|
Examples of NON row echlon form:
|
|
|
|
\begin{equation*}
|
|
\begin{amatrix}{2}
|
|
0 & \textbf{1} & 0 \\
|
|
\textbf{1} & 0 & 0 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{2}
|
|
0 & 0 & \textbf{1} \\
|
|
\textbf{1} & -1 & 0 \\
|
|
0 & 0 & 1 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{3}
|
|
\textbf{1} & 0 & 2 & 1 \\
|
|
0 & \textbf{1} & 0 & 2 \\
|
|
0 & \textbf{1} & 1 & 3 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{4}
|
|
\textbf{0} & \textbf{0} & \textbf{0} & \textbf{0} & \textbf{0} \\
|
|
1 & 0 & 2 & 0 & 1 \\
|
|
0 & 0 & 0 & 1 & 3 \\
|
|
0 & 0 & 0 & 0 & 0 \\
|
|
\end{amatrix}
|
|
\end{equation*}
|
|
\subsection{Reduced Row-Echelon Form}
|
|
|
|
\begin{defn}[Reduced Row-Echelon Form]
|
|
Suppose an augmented matrix is in row-echelon form. It is in \textbf{reduced row-echelon form} if
|
|
\begin{itemize}
|
|
\item Leading entry of every nonzero row is 1
|
|
\subitem Every pivot point is one
|
|
\item In each pivot column, except the pivot point, all other entries are 0.
|
|
\end{itemize}
|
|
\end{defn}
|
|
|
|
Examples of reduced row-echelon form:
|
|
|
|
\begin{equation*}
|
|
\begin{amatrix}{2}
|
|
1 & 2 & 3 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{2}
|
|
0 & 0 & 0 \\
|
|
0 & 0 & 0 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{2}
|
|
1 & 0 & 0 \\
|
|
0 & 1 & 0 \\
|
|
0 & 0 & 1 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{3}
|
|
1 & 0 & 0 & 1 \\
|
|
0 & 1 & 0 & 2 \\
|
|
0 & 0 & 1 & 3 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{4}
|
|
0 & 1 & 2 & 0 & 1 \\
|
|
0 & 0 & 0 & 1 & 3 \\
|
|
0 & 0 & 0 & 0 & 0 \\
|
|
0 & 0 & 0 & 0 & 0 \\
|
|
\end{amatrix}
|
|
\end{equation*}
|
|
|
|
Examples of row-echelon form but not reduced: (pivot point is not 1 / all other elements \textbf{in pivot column} must be zero)
|
|
|
|
\begin{equation*}
|
|
\begin{amatrix}{2}
|
|
\textbf{3} & 2 & 1 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{2}
|
|
1 & \textbf{-1} & 0 \\
|
|
0 & 1 & 0 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{2}
|
|
\textbf{2} & 0 & 0 \\
|
|
0 & 1 & 0 \\
|
|
0 & 0 & 1 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{3}
|
|
\textbf{-1} & 2 & 3 & 4 \\
|
|
0 & 1 & 1 & 2 \\
|
|
0 & 0 & 2 & 3 \\
|
|
\end{amatrix}
|
|
\begin{amatrix}{4}
|
|
0 & 1 & 2 & \textbf{8} & 1 \\
|
|
0 & 0 & 0 & \textbf{4} & 3 \\
|
|
0 & 0 & 0 & 0 & 0 \\
|
|
0 & 0 & 0 & 0 & 0 \\
|
|
\end{amatrix}
|
|
\end{equation*}
|
|
|
|
To note: 2nd matrix has -1 in the pivot column, but 5th matrix has 2 in a non-pivot column so its fine
|
|
|
|
\subsection{Solving Linear System}
|
|
|
|
If Augmented Matrix is in reduced row-echelon form, then solving it is easy
|
|
|
|
\begin{equation*}
|
|
\begin{amatrix}{3}
|
|
1 & 0 & 0 & 1 \\
|
|
0 & 1 & 0 & 2 \\
|
|
0 & 0 & 1 & 3 \\
|
|
\end{amatrix}
|
|
\text{then } x_1 = 1, x_2 = 2, x_3 = 3
|
|
\end{equation*}
|
|
|
|
\begin{note}
|
|
\begin{itemize}
|
|
\item If any equations in the system is inconsistent, the whole system is inconsistent
|
|
\end{itemize}
|
|
\end{note}
|
|
|
|
\subsubsection{Examples}
|
|
|
|
Augmented Matrix: $\begin{amatrix}{4}
|
|
1 & -1 & 0 & 3 & -2 \\
|
|
0 & 0 & 1 & 2 & 5 \\
|
|
0 & 0 & 0 & 0 & 0 \\
|
|
\end{amatrix}$
|
|
|
|
\begin{itemize}
|
|
\item The zero row can be ignored.
|
|
\systeme{x_1 - x_2 + 3x_4 = -2, x_3 + 2x_4 = 5}
|
|
\item Degree of freedom(\# cols): 4, number of restrictions (\# pivot cols): 2, arbitrary vars(\# non pivot cols): 4-2 = 2. Set this to the non-pivot cols
|
|
\end{itemize}
|
|
\begin{enumerate}
|
|
\item Let $x_4 = t$ and sub into 2nd eqn
|
|
\subitem $x_3 + 2t = 5 \Rightarrow x_3 = 5-2t$
|
|
\item sub $x_4 = t$ into 1st eqn
|
|
\subitem $x_1 - x_2 + 3t = -2$
|
|
\subitem Let $x_2 = s$. Then $x_1 = -2 + s - 3t$
|
|
\item Infinitely many sols with ($s$ and $t$ as arbitrary params)
|
|
\subitem $x_1 = -2 + s - 3t, x_2 = s, x_3 = 5-2t, x_4 = t$
|
|
|
|
\end{enumerate}
|
|
|
|
Augmented Matrix: $\begin{amatrix}{5}
|
|
0 & 2 & 2 & 1 & -2 & 2 \\
|
|
0 & 0 & 1 & 1 & 1 & 3 \\
|
|
0 & 0 & 0 & 0 & 2 & 4 \\
|
|
\end{amatrix}$
|
|
\begin{itemize}
|
|
\item \systeme{0x_1 + 2x_2 + 2x_3 + 1x_4 -2x_5 = 2,x_3 + x_4 +x_5 = 3,2x_5 = 4}
|
|
\item Degree of freedom: 5, number of restrictions: 3, arbitrary vars: 5-3 = 2
|
|
\end{itemize}
|
|
|
|
\begin{enumerate}
|
|
\item by 3rd eqn, $2x_5 = 4 \Rightarrow x_5 = 2$
|
|
\item sub $x_5 = 2$ into 2nd eqn
|
|
\subitem $x_3 + x_4 + 2 = 3 \Rightarrow x_3 + x_4 = 1$
|
|
\subitem let $x_4 = t$. Then $x_3 = 1-t$
|
|
\item sub $x_5 = 2, x_3 = 1-t, x_4 = t$ into 1st eqn
|
|
\subitem $2x_2 + 2(1-t) + t - 2(2) = 2 \Rightarrow 2x_2 -t = 4 \Rightarrow x_2 = \frac{t}{2} + 2$
|
|
\item system has inf many solns: $x_1 = s, x_2 = \frac{t}{2} + 2, x_3 = 1-t, x_4 = t, x_5 = 2$ where $s$ and $t$ are arbitrary
|
|
\end{enumerate}
|
|
|
|
\subsubsection{Algorithm}
|
|
Given the augmented matrix is in row-echelon form.
|
|
\begin{enumerate}
|
|
\item Set variables corresponding to non-pivot columns to be arbitrary parameters
|
|
\item Solve variables corresponding to pivot columns by back substitution (from last eqn to first)
|
|
\end{enumerate}
|
|
|
|
|
|
\subsection{Gaussian Eliminiation}
|
|
|
|
\begin{defn}[Gaussian Elimination]\ \\
|
|
\begin{enumerate}
|
|
\item Find the left most column which is not entirely zero
|
|
\item If top entry of such column is 0, replace with nonzero number by swapping rows
|
|
\item For each row below top row, add multiple of top row so that leading entry becomes 0
|
|
\item Cover top row and repeat to remaining matrix
|
|
\end{enumerate}
|
|
\end{defn}
|
|
|
|
\begin{note}[Algorithm with Example]\ \\
|
|
$\begin{amatrix}{6}
|
|
0 & 0 & 0 & 2 & 4 & 2 & 8 \\
|
|
0 & 1 & 2 & 4 & 5 & 3 &-9 \\
|
|
0 &-2 &-4 &-5 &-4 & 3 & 6 \\
|
|
\end{amatrix}$
|
|
\begin{enumerate}
|
|
\item Find the left most column which is not all zero (2nd column)
|
|
\item Check top entry of the selection. If its zero, replace it by a nonzero number by interchanging the top row with another row below
|
|
\subitem $\begin{amatrix}{6}
|
|
0 & 1 & 2 & 4 & 5 & 3 &-9 \\
|
|
0 & 0 & 0 & 2 & 4 & 2 & 8 \\
|
|
0 &-2 &-4 &-5 &-4 & 3 & 6 \\
|
|
\end{amatrix}$
|
|
\item For each row below the top row, adda suitable multiple of top row so that leading entry becomes 0.
|
|
\subitem $2R_1 + R_3$ will ensure that the -2 turns to 0
|
|
\subitem $\begin{amatrix}{6}
|
|
0 & 1 & 2 & 4 & 5 & 3 &-9 \\
|
|
0 & 0 & 0 & 2 & 4 & 2 & 8 \\
|
|
0 & 0 & 0 & 3 & 6 & 9 &-12\\
|
|
\end{amatrix}$
|
|
\item Cover top row and repeat procedure to the remaining matrix
|
|
\subitem $\begin{amatrix}{6}
|
|
0 & 1 & 2 & 4 & 5 & 3 &-9 \\
|
|
\hline
|
|
0 & 0 & 0 & 2 & 4 & 2 & 8 \\
|
|
0 & 0 & 0 & 3 & 6 & 9 &-12\\
|
|
\end{amatrix}$
|
|
\subitem Look at $C_4$. $R_3 \times -1.5R_2$ will set $R_3C_4$ to zero.
|
|
\subitem $\begin{amatrix}{6}
|
|
0 & 1 & 2 & 4 & 5 & 3 &-9 \\
|
|
\hline
|
|
0 & 0 & 0 & 2 & 4 & 2 & 8 \\
|
|
0 & 0 & 0 & 0 & 0 & 6 & -24\\
|
|
\end{amatrix}$
|
|
\subitem This is now in row echelon form.
|
|
\end{enumerate}
|
|
Only use $R_i \Leftrightarrow R_j or R_i + CR_j$ in this method.
|
|
\end{note}
|
|
|
|
\subsection{Gauss-Jordan Elimination}
|
|
\begin{defn}[Gauss Joran Elimination]\ \\
|
|
\begin{enumerate}
|
|
\item[1-4.] Use Gaussian Eliminiation to get row-echelon form
|
|
\setcounter{enumi}{4}
|
|
\item For each nonzero row, multiply a suitable constant so pivot point becomes 1
|
|
\item Begin with last nonzero row and work backwords
|
|
\subitem Add suitable multiple of each row to the rows above to introduce 0 above pivot point
|
|
\end{enumerate}
|
|
\begin{itemize}
|
|
\item Every matrix has a unique reduced row-echelon form.
|
|
\item Every nonzero matrix has infinitely many row-echelon ofrm
|
|
\end{itemize}
|
|
\end{defn}
|
|
\begin{note}[Gauss Jordan Elimination Example] Suppose an augmented matrix is in row-echelon form.
|
|
$\begin{amatrix}{5}
|
|
1 & 2 & 4 & 5 & 3 & -9 \\
|
|
0 & 0 & 2 & 4 & 2 & 8 \\
|
|
0 & 0 & 0 & 0 & 6 & -24 \\
|
|
\end{amatrix}$
|
|
\begin{enumerate}
|
|
\item All pivot points must be 1
|
|
\subitem multiply $R_2$ by $\frac{1}{2}$ and $R_3$ by $\frac{1}{6}$
|
|
\subitem $\begin{amatrix}{5}
|
|
1 & 2 & 4 & 5 & 3 & -9 \\
|
|
0 & 0 & 1 & 2 & 1 & 4 \\
|
|
0 & 0 & 0 & 0 & 1 & -4 \\
|
|
\end{amatrix}$
|
|
\item In each pivot col, all entries other than pivot point must be 0. Work backwards
|
|
\subitem $R_1 + -3R_1$, $R_2 + -R_1$
|
|
\subitem $\begin{amatrix}{5}
|
|
1 & 2 & 4 & 5 & 0 & 3 \\
|
|
0 & 0 & 1 & 2 & 0 & 8 \\
|
|
0 & 0 & 0 & 0 & 1 & -4 \\
|
|
\end{amatrix}$
|
|
\subitem $R_1 + -4R_2$
|
|
\subitem $\begin{amatrix}{5}
|
|
1 & 2 & 0 & -3 & 0 & -29 \\
|
|
0 & 0 & 1 & 2 & 0 & 8 \\
|
|
0 & 0 & 0 & 0 & 1 & -4 \\
|
|
\end{amatrix}$
|
|
|
|
\end{enumerate}
|
|
\end{note}
|
|
|
|
|