$a)$ Let $a$ be a (column) vector in $\mathbb{R}^n$.

Show that the function $f(x)=a^Tx$ is linear function from $\mathbb{R}^n$ to $\mathbb{R}$

$b)$ Show that any linear function $f:\mathbb{R}^n\to\mathbb{R}$ is of the form $f(x)=a^Tx$ for some vector $a\in\mathbb{R^n}$.

Hint let $e_i$ denote the standard unit vectors in $\mathbb{R}^n$ and write $x$ as $x=x_1e_1+\dots e_nx_n.$ Let $a_i:=f(e_i)\in\mathbb{R}.$ What is $f(x_1e_1+\dots+e_nx_n)$

I proved $a)$.



$$\text{WTS } f \text{ is a function s.t. } f:\mathbb{R}^n\to\mathbb{R}\text{ and }f \text{ is linear}$$

First part:

By given info that

$$a:=\begin{pmatrix}a_1\\\vdots\\a_n\end{pmatrix},x:=\underbrace{\begin{pmatrix}x_1\\\vdots\\x_n\end{pmatrix}}_{\text{input}\in\mathbb{R}^n},a^T=\begin{pmatrix}a_1\\\vdots\\a_n\end{pmatrix}^T=\begin{pmatrix}a_1\space\dots\space a_n\end{pmatrix}$$

$$\Rightarrow a^Tx=\begin{pmatrix}a_1\space\dots\space a_n\end{pmatrix}\begin{pmatrix}x_1\\\vdots\\x_n\end{pmatrix}=\underbrace{\sum_{i=1}^na_ix_i}_{\text{output}\in\mathbb{R}}$$

So we indeed have $\boxed{f:\mathbb{R}^n\to\mathbb{R}}$

Second part:

To show $f$ is linear

Let $$x_1:=\begin{pmatrix}x_{1_1}\\\vdots\\x_{1_n}\end{pmatrix},x_2:=\begin{pmatrix}x_{2_1}\\\vdots\\x_{2_n}\end{pmatrix},x:=cx_1+x_2=\begin{pmatrix}cx_{1_1}+x_{2_1}\\\vdots\\cx_{1_n}+x_{2_n}\end{pmatrix}$$

(To avoid confusion, notice $x_1,x_2,x$ are redefined here, which is different from First part)

Just verify that $$f(cx_1+x_2)=cf(x_1)+f(x_2)$$






Be honestly I'm not sure I understood what $b)$ is asking, here is my attempts

My thoughts

Try to formalize 'any linear function', but what does this term really mean? For example $f_1(x)=x+1$ and $f_2(x)=x+2-1$, are they the same linear function?

Maybe we use another definition: $$\text{two functions with same domain are different iff they have different range}$$

So the proof should show that each functons with domain $\mathbb{R}^n$ that are different can all be written in form of $a^Tx$, in another words the range of each different $f$ in this form biject to all the possible subsets of $\mathbb{R}$.

Start from the hint.


Let $$e_1\dots e_n\text{, be all standard basis of }\mathbb{R}^n$$

$$a_i:=f(e_i)\in\mathbb{R} \text{, where } i\in[1,n]\cap\mathbb{Z}$$

That $f(x)=f(x_1e_1+\dots+e_nx_n)=x_1f(e_1)+\dots +x_nf(e_n)=a_1x_1\dots a_nx_n=a^Tx\tag*{$\square$}$


For $b)$, suppose $f:\mathbb{R}^n\to\mathbb{R}$ is a linear transformation. We need to show that $f$ is necessarily in the form $f(x)=a^tx$ for some $a\in\mathbb R^n$. Since any linear map is determined by what it does to the elements of the basis, we can define $a_i:=f(e_i)$ for $i=1,2,\dots,n$.

Then, for all $x=(x_1,x_2,\dots,x_n)\in\mathbb{R}^n$ note the following $$\begin{align} f(x) &= f\left( \sum_{i=1}^n x_ie_i \right) = \sum_{i=1}^n x_if(e_i) = \sum_{i=1}^n a_ix_i \\ &= \begin{pmatrix} a_1&a_2&\cdots&a_n \end{pmatrix} \begin{pmatrix} x_1\\x_2\\ \vdots \\ x_n \end{pmatrix} = a^tx. \end{align}$$ here, $a=(a_1,a_2,\dots,a_n)^t$, obviously.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.