Basis for set of vectors is wrong

80 Views Asked by At

Original matrix A is:

$$\begin{pmatrix} 1 & 2 & 7& 1\\ 1 & 1 & 2& 0\\ 2&0&-6&-2 & \\ 1&1&2&0 \end{pmatrix}$$

$w=(A*\begin{pmatrix} x1\\ x2\\ x3\\ x4 \end{pmatrix}=\begin{pmatrix} 0\\ 0\\ 0\\ 0 \end{pmatrix})$ Find base set of vector for W.

Using Gaus method I get:

$$\begin{pmatrix} 1 & 2 & 7&1\\ 0&-1&-5&-1 \\ 0 & 0 & 0&0\\ 0&0&0&0 \end{pmatrix}$$ which has 2 leading elements. So the column of each elemnt suppose to be part of base, these are: $(1,1,2,1)$ , $(2,1,0,1)$ but seemingly it's wrong because the answer is written like this: ${(3,-5,1,0),(1,-1,0,1)}$

where is my mistake?

2

There are 2 best solutions below

0
On

Take your reduced system

$$\begin{pmatrix}1 &2 &7& 1\\0 &-1 &-5 &-1\\ 0 &0 &0 &0\\ 0 &0 &0 &0\end{pmatrix}\begin{pmatrix}x\\y\\z\\w\end{pmatrix}=\begin{pmatrix}x+2y+7z+w\\-y-5z-w\\0\\0\end{pmatrix}=\begin{pmatrix}0\\0\\0\\0\end{pmatrix}$$

You now just have to solve the equation for $v=(x,y,z,w)^\top$. Working with the method of back-substitution, we first assign all the non-pivots(the variables not leading any row) a real constant, as they may vary in their value arbitrarily and we then can determine the pivot variables in dependence on them.

Thus, let $w=c,z=c'$(for constants $c,c'$). Then, by the second row of the vector equation, we may determine the value of $y$ in terms of the constants $c,c'$(values of $z,w$):

$$-y-5z-w=0\Leftrightarrow-y-5c'-c=0\Leftrightarrow y=-5c'-c$$

Continuing back-substitution, we now derive from the first row the value of $x$ in dependence of $c,c'$:

$$x+2y+7z+w=0\Leftrightarrow x=-2y-7z-w\Leftrightarrow x=10c'+2c-7c'-c\Leftrightarrow x=3c'+c$$

Thus, your solutions are any vectors of the following form

$$(3c'+c,-5c'-c,c',c)=c'(3,-5,1,0)+c(1,-1,0,1)$$

for arbitrary $c,c'\in\mathbb{R}$.


The solution set of $Av=\mathbf0$, i.e. the set $S(A,\mathbf0)=\{v\in\mathbb{R}^4\mid Av=\mathbf0\}$, for which you want to determine a basis, is commonly called the kernel of the matrix $A$, written $\mathrm{ker}(A)$.

By the above, we've found that any solution to $Av=\mathbf0$, i.e. any $v\in\mathrm{ker}(A)$, can be written as

$$v=c'(3,-5,1,0)+c(1,-1,0,1)$$

Thus, in other terminology, we say that any $v\in\mathrm{ker}(A)$ is a linear combination of $(3,-5,1,0)$ and $(1,-1,0,1)$. The set of all linear combination for a set of vectors $S$ is called the span(or linear hull) of the set $S$ and I denote this with $\mathrm{span}(S)$.

We may thus rephrase our result about the solutions of $\mathrm{ker}(A)$ as such, that

$$\mathrm{ker}(A)=\mathrm{span}(\{(3,-5,1,0),(1,-1,0,1)\})$$

A basis, for any space $V$, is defined as linearly independent set of vectors spanning $V$, i.e. a linearly independent set $B$, s.t. $\mathrm{span}(B)=V$.

Now, as $(3,-5,1,0),(1,-1,0,1)$ are linearly independent(check this yourself), it follows that they also constitute a basis for this set(space).


Maybe, on a last note, try to generalize the process I've just outline. Toying with these notions and especially with the connections between subspaces, linear systems of equation and matrices in combination is a great exercise. E.g. show that for any $n\times n$ real matrix $A$, $\mathrm{ker}(A)$ always forms a subspace. Is this true for any solution set $S(A,b)$ for a vector equation $Av=b$?

0
On

There is a standard method for finding a basis for the null space of a matrix. Compute the reduced row echelon form (RREF): \begin{align} A=\begin{pmatrix} 1 & 2 & 7 & 1\\ 1 & 1 & 2 & 0\\ 2 & 0 &-6 &-2 \\ 1 & 1 & 2 & 0 \end{pmatrix} &\to \begin{pmatrix} 1 & 2 & 7 & 1\\ 0 &-1 &-5 &-1\\ 0 &-4 &-20 &-4 \\ 0 &-1 &-5 & -1 \end{pmatrix} &&\begin{aligned} R_2&\gets R_2-R_1 \\ R_3&\gets R_3-2R_1 \\ R_4&\gets R_4-R_1\end{aligned} \\[6px] &\to \begin{pmatrix} 1 & 2 & 7 & 1\\ 0 & 1 & 5 & 1\\ 0 &-4 &-20 &-4 \\ 0 &-1 &-5 & -1 \end{pmatrix} && R_2\gets -R_2 \\[6px] &\to \begin{pmatrix} 1 & 2 & 7 & 1\\ 0 & 1 & 5 & 1\\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \end{pmatrix} &&\begin{aligned} R_3&\gets R_3+4R_2 \\ R_4&\gets R_4+R_2\end{aligned} \\[6px] &\to \begin{pmatrix} 1 & 0 &-3 &-1\\ 0 & 1 & 5 & 1\\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \end{pmatrix} && R_1\gets R_1-2R_2 \end{align} The system $Ax=0$ is equivalent to the system $Ux=0$, where $U$ is the last matrix above (the RREF of $A$). This system can be rewritten as \begin{cases} x_1 = 3x_3+x_4 \\[4px] x_2 = -5x_3-x_4 \end{cases} and a basis for the null space can be obtained by choosing first $x_3=1$ and $x_4=0$, then $x_3=0$ and $x_4=1$, yielding the vectors $$ \begin{pmatrix} 3 \\ -5 \\ 1 \\ 0 \end{pmatrix} \qquad\text{and}\qquad \begin{pmatrix} 1 \\ -1 \\ 0 \\ 1 \end{pmatrix} $$