Monday September 20 − Lecture 4 : Projections to lines in R3 (Refers to section 4.
2)
Expectations:
1. Graphically represent the projection of a vector b onto a vector a (or onto a line L)
in R2 and R3.
2. Find the projection of a vector onto another vector or line in R2 and R3.
3. Graphically represent the vector w = perpab.
4. Find the distance between a point b and a line L in R3.
5. Find the distance between a point b and a plane P(n, a) in R3.
4.1 Projections − Occasionally, one might want to express a vector x as a sum of two
orthogonal vectors (perpendicular vectors), called orthogonal component vectors of x.
Suppose we are given a vector b in R2 and another vector a. Suppose we want to express
b as the sum of a scalar multiple of a, say ta, and another vector w, such that ta ⊥ w. In
this lecture we study how this can be done.
4.2 Definition − Let b be a vector in R2 or R3 and let a be any other vector in the same
space. We define the projection of b onto a as being the vector p = proja(b) = ta, where t
is such w = b − p is orthogonal to p.
• So vector p = proja(b) is a scalar multiple of the directed line segment a.
• If in R2, we have defined p = proja(b) so that the points (0, 0), p, and b form a
right triangle. (The directed line segment of b forming the hypotenuse).
• The expression is b − proja(b) denoted by perpa(b) . (Pronounced “perp b onto
a”)
• So it is always true that
b = proja(b) + perpa(b)
4.2.1 Example − Consider the vector b = (−4, 3) and let a = (2, 0). It is easy, in this
case, to graphically visualize, and then compute what proja(b) is.
• Then proja b = −2(2, 0) = (−4, 0) and w = (−4, 3) − (−4, 0) = (0, 3).
4.3 We will show that the projection of b onto a is given by the expression
• Suppose p = ta and the scalar t is such that w = b − p is orthogonal to p. Then
< ta, b − ta > = 0.
• Hence
4.3.1 Example − Find the projection of u = (2, 3, 1) onto v = (1, 2, −6) . Also find
the norm of the projection
• We want proj(1, 2, -6)(2, 3, 1). Note that || v ||2 = 41 and < u, v > = 2. Hence
proj(1, 2, −6)(2, 3, 1) = [ < u, v > ] ( v / || v ||2 ) = (2/41)(1, 2, −6).
• The length of the projection is
|| proj(1, 2, -6)(2, 3, 1) || = | < u, v > | / || v || = 2 / √41
4.3.2 Example – If b is the point (2, 2, 2) and L is the line t(1, 0, 1) find the point on L
which is closest to b.
• The point on L closest to b is the projection of b onto the direction vector
d = (1, 0, 1) of L.
• Verify that projdb = (2, 0, 2)
4.4 The shortest distance from a point b to a line a + td in R3 .
Suppose we are given a line L, a + td, and a vector b in R3 and we wish to find the point
on L which is closest to b and the shortest distance from b to the line L.
• In fact we want to find the length of the line segment which links b
perpendicularly to L.
• To do this we find the projection p of b − a onto d .
• The point on the L which is closest to b is p + a.
• The shortest distance from b to L is || b − (p + a) || or equivalently || (b − a) − p ||
4.4.1 Example − Let L be a line passing through the points (1, 2, 3) to (1, 5, 7). Find
the point on the line L which is closest to the point b = (−3, 0, 4). What is this
distance.
• We first obtain the equation of the line: The direction vector of L can be chosen as d =
(1, 5, 7) − (1, 2, 3) = (0, 3, 4). So L : a + td = (1, 2, 3) + t(0, 3, 4).
• We now find p = projd b − a.
• p = < (−4, −2, 1) , (0, 3, 4) > (1/25)(0, 3, 4) = (−2/25)(0, 3, 4)
• The point on L closest to b is p + a = (−2/25)(0, 3, 4) + (1, 2, 3).
|| b − (p + a) || = 4.57 (check!)
4.5 The shortest distance of a vector b to a plane P(n, a) in R3 .
The shortest distance of a vector b to a plane P(n, a) is the length of the projection of b −
a onto the normal n of the plane:
4.5.1 Example − Find the shortest distance between the point b = (−3, 0, 4) and the
plane 3x + 4y – z = 2. Find the point on the plane which is closest to the vector b.
• Choose the vector a = (0, 0, −2) on the plane 3x + 4y – z = 2. A normal vector of
the plane is n = (3, 4, −1).
• We first project the vector b – a = (−3, 0, 4) – (0, 0, −2) = (−3, 0, 6) onto the
normal vector n = (3, 4, −1). The shortest distance is the length of projn(b – a).
• We get as distance
• We now seek the point x the plane closest to b:
We find x from b – p:
Recommended exercises: §4.2 : 11, 12, 21, 23 a), 24 a) 25, 39, 44 a)
