11

11.1

Limits of Functions of Real Variables

The denitions and theorems for limits of sequences can easily be extended to limits of functions as x ; we can simply replace the integer variable n with the real variable x, and everything that we have discussed so far works exactly the same way. We can even adapt it all for limits as x ; nothing else changes. However, given a function f of a real variable x, we can also consider a second type of limit: given a constant a, what happens to f (x) as x a? This requires a new denition: Denition: The statement f (x) L as x a or lim f (x) = L means that for any positive number , there exists a number such that
x a

0 < | x a| <

| f ( x) L| < .

Whereas for sequences we needed to make n large, were now trying to make the distance between x and a small ; you should think of as being a small positive number. If you found the idea of the game from the previous section helpful, then the modication is this: if I suggest a small value for , your challenge is now to tell me how close x must be to a in order for |f (x) L| to be less than . The number you provide is what weve labelled as . If you can always win this game, then the limit is indeed L. Theres one other small modication: weve stipulated that 0 < |x a| because were not interested in what happens when x is exactly equal to a. The denition could be paraphrased as lim f (x) = L means that f (x) can be made arbitrarily close to L by making x suciently close (but not equal) to a. Example: Prove that 3x 2 4 as x 2. Solution: We must show that for any > 0, there exists a > 0 such that |x 2| < = |(3x 2) 4| < . Well, working backwards again,
x a

|(3x 2) 4| <

= |3x 6| < = 3 |x 2| < = |x 2| < . 3 We can now see that = /3 is the value required for the proof: (Condensed) Proof: Let be any positive number. If |x 2| < /3, then |(3x 2) 4| < . Note that limits of functions can also fail to exist in the same ways as limits of functions can fail to exist. Also note that we need a dierent denition for every one of the following statements: lim f (x) = L
x a x a x a

lim f (x) = lim f (x) = lim f (x) = L

lim f (x) = lim f (x) =

lim f (x) = L lim f (x) = lim f (x) =

You may get some practice with some of these in the assignments. Actually, there are six more; we also need denitions of one-sided limits, L lim f (x) = x a + L lim f (x) = x a

and

11.2

Calculation of Limits (of Functions of Real Variables)

Fortunately, the denition weve just introduced leads to the same set of theorems as govern limits of sequences as n (and functions as x ). We also get one more extremely important one: f is continuous at a value a if and only if f (x) f (a) as x a. (1)

If you prefer the other notation, this says that f is continuous at a if and only if lim f (x) = f ( a) .
x a

This theorem makes the calculation of most limits trivial in practice. For instance, consider the example above. It seems silly to have to prove that 3x 2 approaches 4 as x approaches 2, because we all know that the function 3x 2 is continuous (not just at 2, but everywhere), so all we have to do is evaluate it at x = 2! In fact, we know that all polynomials are continuous, and all of our other familiar functions are continuous on their domains as well. Therefore evaluating limits only requires real eort when we encounter discontinuities.1 So, what kinds of discontinuities might we encounter? The most common is the divisionby-zero kind. If both numerator and denominator approach zero as x a, then we have an indeterminate form, and as we saw with sequences the most useful strategy is to try to rewrite the function in a form which is not indeterminate. In some cases, though, it isnt possible to rewrite the function in any useful way, so we might ask ourselves if the Squeeze Theorem could be of any use. We may also encounter discontinuities when working with piecewise-dened functions. For these we need one more (hopefully transparent) theorem: f (x) L as x a if and only if f (x) L as x a+ and f (x) L as x a . Here the notations x a+ and x a refer to the limits of f as we approach a from the right and left sides, respectively (and again, all of our other limit theorems extend to these
Technically, we have a serious logical problem here. We have not yet dened continuity! What we should be doing is proceeding to a discussion of continuity (which well discuss in the next lecture), verifying this claim that all of our elementary functions are continuous on their domains, and then returning to the present discussion of limits. However, we all have a basic understanding of what continuity means, and this departure from the logical procession allows us proceed to some examples of actual calculations. The diculty is this: an understanding of limits is required for a rigorous discussion of continuity, but in practice we rely on an understanding of continuity to evaluate limits! Weve dealt with this in a dierent way than most textbooks, but the textbooks still have a logical problem. They use theorem (1) as the denition of continuity... but then they assume that their functions are continuous almost everywhere in order to determine if f (x) f (a)!
1

types of limits as well). If we nd dierent limits from the two directions of approach, then f simply does not have a limit as x approaches a. So, for piecewise-dened functions, we simply need to check whether the two limits match! Examples: a) lim
x 2 x 2

This function is continuous at x = 2, so we can conclude immediately that 2x 2 + 1 2x2 + 1 9 3 lim 2 = 2 = = . x 2 x + 6 x 4 x + 6x 4 x=2 12 4

2x2 + 1 + 6x 4

b) lim

x 2

This time we have a discontinuity in the denominator, but we do not have an indeterminate form, since the numerator is not zero. A little bit of thought is all that is required x+5 here: if we approach 2 from the right we obtain large positive values so lim = , x 2 + x 2 x+5 while if we approach from the left we obtain large negative values so lim = . x 2 x 2 Therefore the limit simply doesnt exist. c) lim x2 + x 6 x2

x+5 x2

x 2

Here we do have an indeterminate form; both numerator and denominator approach x2 + x 6 (x + 3)(x 2) zero as x 2, so we try rewriting the function: = =x+3 x2 x2 (for x = 2). Clearly x + 3 5 as x 2, so the limit is 5. d) lim
x 0

(2 + x)3 8 x

Again, a little bit of algebra is all thats required: for x = 0, (2 + x)3 8 x3 + 6x2 + 12x = = x2 + 6x + 12, x x

and this approaches 12 as x 0. 1 e) lim x cos x 0 x 1 This is a tricky one; cos x behaves very strangely near x = 0 (we may discuss this in 4

class). However, it is bounded: 1 1 cos 1. x Therefore we can state that 1 x x cos x, x

and so we can conclude from the Squeeze Theorem that x cos f) lim 2x 2 3x |2x 3|

1
x

0 as x 0.

x 1 .5

This is a piecewise-dened function, with the change in the denition of the function occurring precisely at the discontinuity, so we need to check the two one-sided limits: For x > 1.5, 2x 2 3x x(2x 3) = = x 1.5 as x 1.5+ . |2x 3| 2x 3 2x 2 3x x(2x 3) = = x 1.5 as x 1.5 . |2x 3| (2x 3)

For x < 1.5,

Since the left- and right-sided limits dont match, the limit does not exist. g) lim 9x 2 + x 3x

Limits as x work exactly as limits of sequences; weve seen problems like this before! 9x 2 + x 3x = 9x2 + x 3x 9x2 + x + 3x 9x 2 ( 9x2 + x 3x)

x + x + 3x

1 = 1 9+ x +3 h)
x

1 as x . 6

lim

x 4x 2 + 3

Limits as x also work the same way... except that occasionally the fact that we

are dealing with negative values of x is important. x 4x2 + 3 1 4x2 + 3 1 = 1 4x 2 + 3 x2 1 , = 3 4+ x 2 =

since x = x2 when x < 0(!!!) which 1 2 as x .

A Special Limit: Once in a while we do encounter limits which we cannot evaluate with these simple algebraic manipulations. In Math 119 well introduce a couple of other techniques which will help (Taylor Series approximations, and LHospitals Rule). For now, well introduce just one special limit which is of particular importance. It can be established with a geometric argument, but well simply state it for reference: sin = 1.

lim

The proof is somewhat long, so well omit it, but we can make a quick intuitive argument as to why this should be true. Recall that is not just the angle, but also the length of the arc subtended by the angle in a unit circle. If we consider an extremely small angle, we can see that this arc becomes virtually indistinguishable from the vertical distance sin (see Figure 1).

Figure 1: sin . You should be able to see that this is dierent; its zero, by the Squeeze Theorem. As a nal note, we point out that the function You may also encounter the limit lim

f ( x) =

sin x
x

if x = 0 if x = 0

occurs so often in digital signal processing that it gets a special name: it is called the sinc function (short for sinus cardinalus, or cardinal sine) , sinc (x). You can see the two limits weve just mentioned in its graph:

0.75

0.5

0.25

-15!

-10!

-5!

5!

10!

15!

-0.25

Figure 2:

-0.5