Let ℓ1 and ℓ2 be lines given in vector form by ℓ1: (-5, -1, 1)+t(2; -3; 1) and ℓ2 : (1; 3; 1) + t(0; 0; 5) respectively. There exists precisely one line in R3 which intersects and is perpendicular to both of these. Find this line in vector form.
the cross product is the vector perpendicular to both lines. the problem is to find the one of the two points intersected by the required line.
oh cross product :P yah that makes sense. hmm
@SithsAndGiggles , any suggestion on finding a particular point?
If the lines \(l_1\) and \(l_2\) have exactly one line which intersects and is perp to both of them, then should \(l_1\) and \(l_2\) be skew?
@Erin001001 , not necessarily. the two lines may be coplanar. but for the given lines, they are skew.
Ah yes, I meant if they didn't intersect (I'd already checked for that).
I think I have a highly tedious way to do this. Using cross product, we can find a vector perp to both \(l_1\) and \(l_2\). Now divide out its length, giving a unit vector \(\underline{n}\) in that direction. Now find the distance \(d\) between the two skew lines and multiply \(\underline{n}\) by \(d\). Let \(\underline{v}=d\underline{n}\). Now there should be a unique point on \(l_1\), say (-5; -1; 1) + s(2; -3; 1), and a unique point on \(l_2\), say (1; 3; 1) + t(0; 0; 5), such that (-5; -1; 1) + s(2; -3; 1) + \(\underline{v}\) = (1; 3; 1) + t(0; 0; 5). (Or perhaps \(-\underline{v}\)?) Solve for s, which gives one of the points of intersection. Can someone check if my method works? Or preferably, come up with a neater one....
|dw:1376878524342:dw|
|dw:1376878609576:dw|
|dw:1376878698753:dw|
\[\vec{PR}=\vec{PQ}\times\vec{RS}=<2,-3,1>\times<0,0,5>\] also, \[\vec{PR}=<6-2t,4+3t,5s-t>\]
now, \[\vec{PR}\cdot\vec{RS}=0\] or \[(6-2t)(0)+(4+3t)(0+(5s-t)(5)=0\] which implies that \[5s=t\]
@sirm3d I'm not sure about this line: \(\vec{PR}=\vec{PQ} \times \vec{RS}=<2,−3,1> \times <0,0,5>\). The length of \(\vec{PQ} \times \vec{RS}\) will give a vector in the direction of \(\vec{PR}\), but the length is necessarily right, surely?
The way I did it was basically to find the length of \(\vec{PR}\) by finding the (shortest) distance between the two skew lines. Then I multiplied it by the unit vector in the \(\vec{PQ} \times \vec{RS}\) direction, which gives \(\vec{PR}\). As this point your method is an improvement on mine. You can just equate \(\vec{PR}\) with \(\vec{PR}=<6−2t,4+3t,5s−t>\), and find s and t. Since we have s, we have one of the intersection points which is what we set out to find.
@Erin001001 , i was doing it your way
Oh. Still it's an improvement: you don't need to deal with the +v or -v nonsense I was going on about...
in the end, we have to solve the equation between \(|\vec{PR}|\) (expressed in terms of t or s) and the actual distance between the skew lines.
Join our real-time social learning platform and learn together with your friends!