23.07.2012 Views

Linear Algebra

Linear Algebra

Linear Algebra

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Topic: Geometry of <strong>Linear</strong> Maps 277<br />

To finish, we will consider a familiar application, in calculus. On the left<br />

below is a picture, like the ones that started this Topic, of the action of the nonlinear<br />

function y(x) =x 2 + x. As described at that start, overall the geometric<br />

effect of this map is irregular in that at different domain points it has different<br />

effects (e.g., as the domain point x goes from 2 to −2, the associated range point<br />

f(x) at first decreases, then pauses instantaneously, and then increases).<br />

5<br />

0<br />

5<br />

0<br />

But in calculus we don’t focus on the map overall, we focus on the local effect<br />

of the map. The picture on the right looks more closely at what this map does<br />

near x =1. Atx = 1 the derivative is y ′ (1) = 3, so that near x =1wehave<br />

that ∆y ≈ 3 · ∆x; in other words, (1.001 2 +1.001) − (1 2 +1)≈ 3 · (0.001). That<br />

is, in a neighborhood of x = 1, this map carries the domain to the codomain by<br />

stretching by a factor of 3 — it is, locally, approximately, a dilation. This shows<br />

a small interval in the domain (x − ∆x..x+∆x) carried over to an interval in<br />

the codomain (y − ∆y..y+∆y) that is three times as wide: ∆y ≈ 3 · ∆x.<br />

x =1<br />

(When the above picture is drawn in the traditional cartesian way then the prior<br />

sentence is usually rephrased as: the derivative y ′ (1) = 3 gives the slope of the<br />

line tangent to the graph at the point (1, 2).)<br />

Calculus considers the map that locally approximates the change ∆x ↦→<br />

3 · ∆x, instead of the actual change map ∆x ↦→ y(1 + ∆x) − y(1), because the<br />

local map is easy to work with. Specifically, if the input change is doubled, or<br />

tripled, etc., then the resulting output change will double, or triple, etc.<br />

5<br />

0<br />

y =2<br />

3(r ∆x) =r (3∆x)<br />

(for r ∈ R) and adding changes in input adds the resulting output changes.<br />

3(∆x1 +∆x2) =3∆x1 +3∆x2<br />

In short, what’s easy to work with about ∆x ↦→ 3 · ∆x is that it is linear.<br />

5<br />

0

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!