3
$\begingroup$

I'm new to Vector and 3D Geometry and today I was taught the equation of a line passing through two points $A(\vec{a})$ and $B(\vec{b})$ as $$\vec{r}=\vec{a} + λ(\vec{b}-\vec{a})$$

It is difficult for me to imagine how this works. Why do we have a $λ$ here? What I can think of is that $\vec{r}=\vec{a} + λ(\vec{b}-\vec{a}$) means that if we draw a line passing through $A$ and $B$ and we plug in different values of $λ$, we will get the position vector of different points that lie on that line. Am i right in thinking so? So this is a bit different from the stuff we use in $2D$ geometry or other forms like expressing the line in terms of $x,y,z$?

$\endgroup$
5
  • $\begingroup$ Yes you are right in your thinking. $\lambda=0$ corresponds to $A(\vec{a})$ and if you move parallel to a vector $\vec{b}$ you will get the line $\vec{r}=\vec{a}+\lambda \vec{b}$ $\endgroup$ Oct 20 at 15:29
  • 2
    $\begingroup$ That equation isn't quite right. $\endgroup$
    – PM 2Ring
    Oct 20 at 15:30
  • 1
    $\begingroup$ @PM2Ring oops, my bad, I did correct it now. $\endgroup$
    – Vega
    Oct 20 at 16:21
  • $\begingroup$ That's much better. :) $\endgroup$
    – PM 2Ring
    Oct 20 at 16:22
  • 1
    $\begingroup$ FWIW, another useful way to write this is $$\vec{r}=(1-\lambda)\vec{a} + \lambda\vec{b}$$ $\endgroup$
    – PM 2Ring
    Oct 20 at 16:35
3
$\begingroup$

Here is a visual explanation, for vectors in $\mathbb{R}^2$. The vectors $\vec{a}$ and $\vec{b}$ fix two points in space. Consider the difference $\vec{b}-\vec{a}$ shown by the red arrows. Then, we move a certain number of lengths of $\vec{b}-\vec{a}$ (given in the picture by $\lambda=5$) along the direction of $\vec{b}-\vec{a}$. The resulting vector $\vec{a}+\lambda(\vec{b}-\vec{a})=\vec{a}+5(\vec{b}-\vec{a})$ has its tip on a line that clearly goes through the tips of $\vec{a}$ and $\vec{b}$ and is parallel to $(\vec{b}-\vec{a})$.

In general, $\lambda$ can be anything, so there can be any number of $\vec{b}-\vec{a}$'s, including fractional amounts, added to $\vec{a}$. This traces out the entire line.

enter image description here

$\endgroup$
4
  • $\begingroup$ +1, Awesome visual representation. So you basically have $\vec{a}$ and then $\vec{b}-\vec{a}$ is a vector parallel to the line joining $A$ and $B$ and this multiplied by $λ$ gives different vectors along the line which when added with $\vec{a}$ by triangular law of addition gives the position vector of different points on the line. Did I get it correct? $\endgroup$
    – Vega
    Oct 20 at 16:26
  • $\begingroup$ @Vega I updated me answer to correspond to your updated post. $\endgroup$
    – march
    Oct 20 at 16:31
  • $\begingroup$ And yes the original (typo) basically corresponds to the equation of a line passing through A and parallel to a vector $\vec{b}$ $\endgroup$
    – Vega
    Oct 20 at 16:37
  • 1
    $\begingroup$ You got it correct. $\endgroup$
    – march
    Oct 20 at 16:38
3
$\begingroup$

Remember, Euclidean vectors are not points; $\begin{pmatrix} 1 \cr 1 \cr 1\end{pmatrix}$ is neither the point $(1,1,1)$ nor affixed between the origin and $(1,1,1);$ rather it is "portable" and represents a particular translation in space.

The parametric equation$$\vec{r}=\vec{a} + λ\vec{u}$$ of a line is actually very intuitive: it is saying that to land on any point $(x,y,z)$ on the line, start at the origin and move directly to point $A$ (after all, $\vec{a}$ is just $A$'s position vector), then move along $\vec{u}$ by a $some$ (negative/zero/positive) multiple of its length.

Then, the position vector $\vec{r}=\begin{pmatrix} x \cr y \cr z\end{pmatrix}$ of that point is given by the vector sum of these two movements/translations.

There is one parameter҂, $\lambda,$ because a line has one degree of freedom҂: the real-valued multiple that I mentioned above is the value of $\lambda$ that lands us on that particular point (and which gives its particular position vector $\vec{r}$). As $\lambda$ varies over $\mathbb R,$ every single point on that line will have been covered.


҂ Lines (including curved lines) in $3$-D space are nonetheless $1$-dimensional objects in the sense that when constrained to the line, a single number (parameter) is sufficient to fully specify the position of each point. This is why their parametric equations (whether in $2$-D space or $3$-D space) contain only a single parameter. Planes, on the other hand, have two parameters in their parametric equation $\vec{r}=\vec{a} + λ\vec{u}+\gamma\vec{v}$.
$\endgroup$
1
  • 1
    $\begingroup$ @RyanG Yeah, either way, in both questions, the concept was the same, which I understood $\endgroup$
    – Vega
    Oct 20 at 17:18
0
$\begingroup$

Rather than looking at a line lets examine a line segment beginning at the point represented by a position vector $p$ and in the direction of some other vector $x$ that tells us the orientation and length of the segment. Geometrically this means we simple add $p + x$ to find where the line segment ends, but what if I want to label every position on the line segment? Well, looking at the half way point along the segment we see that it's at position $p + \frac{1}{2}x$ and a quarter is $p + \frac{1}{4}x$. Examining other points on the segment we see that as long as we pick some scalar multiple of $x$ between zero and one it represents a point on the segment. This gives us the geometric intuition as to why we can now parameterize the line segment as $p + \lambda x$ for some real number $0 \leq \lambda \leq 1$. Now that we have the segment parameterized all we have to do to get the entire line is let $\lambda$ take on any real value. Note that the dimension of the space was irrelevant here as everything was happening in the plane containing $p$ and $x$ so the two dimensional case is sufficient for geometric purposes.

$\endgroup$

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.