Unit 1, lesson 10

Limits at Infinity

What it means for a function to approach a number as x grows without bound, the epsilon-M definition, how leading terms decide the behavior of polynomials and rational functions, and where the number e comes from.

  • Theory part 19
  • Practice part 20
  • Limits at Infinity and Infinite Limits

Theory, part 19

Nothing loads from YouTube until you press play.

Practice, part 20

Nothing loads from YouTube until you press play.

Pour hot coffee into a cup and leave it on the table. Its temperature keeps dropping, but it never goes below room temperature, and after a long time it is as close to room temperature as you care to measure. A limit at infinity describes exactly this, the value a quantity settles toward in the long run.

Key terms

Tending to positive or negative infinity

Definition The notation x→∞x \to \infty means that xx increases without bound, taking values larger than any given number. The notation x→−∞x \to -\infty means xx decreases without bound. Neither ∞\infty nor −∞-\infty is a real number, and xx never equals them. To talk about x→∞x \to \infty, the function must be defined on some interval (c,∞)(c, \infty), and for x→−∞x \to -\infty on some interval (−∞,c)(-\infty, c).

In plain words “As x→∞x \to \infty” means “for larger and larger xx”. It describes a direction, not a place.

Example For f(x)=1/xf(x) = 1/x, the inputs x=10,100,1000x = 10, 100, 1000 give 0.1,0.01,0.0010.1, 0.01, 0.001, and the inputs x=−10,−100,−1000x = -10, -100, -1000 give −0.1,−0.01,−0.001-0.1, -0.01, -0.001. In both directions the outputs close in on 00.

Epsilon-M definition

Definition Let ff be defined on some interval (c,∞)(c, \infty) and let LL be a real number. Then lim⁡x→∞f(x)=L\lim_{x \to \infty} f(x) = L means that for every ε>0\varepsilon > 0 there exists a number MM such that

x>M  ⟹  ∣f(x)−L∣<ε.x > M \implies \lvert f(x) - L \rvert < \varepsilon.

For ff defined on some interval (−∞,c)(-\infty, c), the statement lim⁡x→−∞f(x)=L\lim_{x \to -\infty} f(x) = L means that for every ε>0\varepsilon > 0 there exists MM such that x<Mx < M implies ∣f(x)−L∣<ε\lvert f(x) - L \rvert < \varepsilon.

In plain words Name any tolerance. From some point on, every output stays within that tolerance of LL.

Example To prove that (2x+1)/x→2(2x + 1)/x \to 2 as x→∞x \to \infty, note that for x≠0x \neq 0

∣2x+1x−2∣=1∣x∣.\left\lvert \frac{2x + 1}{x} - 2 \right\rvert = \frac{1}{\lvert x \rvert}.

Given ε>0\varepsilon > 0, take M=1/εM = 1/\varepsilon. If x>Mx > M, then x>0x > 0 and 1/x<1/M=ε1/x < 1/M = \varepsilon.

Finite limits at infinity

Definition When lim⁡x→∞f(x)\lim_{x \to \infty} f(x) or lim⁡x→−∞f(x)\lim_{x \to -\infty} f(x) is a real number LL, it is called a finite limit at infinity. Two basic ones are

lim⁡x→∞1xr=0for every rational r>0,lim⁡x→−∞1xn=0for every positive integer n.\lim_{x \to \infty} \frac{1}{x^r} = 0 \quad \text{for every rational } r > 0, \qquad \lim_{x \to -\infty} \frac{1}{x^n} = 0 \quad \text{for every positive integer } n.

The limit laws hold for limits at infinity as well. The limits at ∞\infty and at −∞-\infty are separate questions, and their answers can differ.

In plain words A function can settle to one value far to the right and a different value far to the left.

Example Consider the function

f(x)=xx2+1.f(x) = \frac{x}{\sqrt{x^2 + 1}}.

For x≠0x \neq 0 we have x2+1=∣x∣1+1/x2\sqrt{x^2 + 1} = \lvert x \rvert \sqrt{1 + 1/x^2}, so

f(x)=x∣x∣⋅11+1/x2,x≠0,f(x) = \frac{x}{\lvert x \rvert} \cdot \frac{1}{\sqrt{1 + 1/x^2}}, \qquad x \neq 0,

and the second factor tends to 11 in both directions. For x>0x > 0 the first factor is 11, and for x<0x < 0 it is −1-1. So

lim⁡x→∞f(x)=1,lim⁡x→−∞f(x)=−1.\lim_{x \to \infty} f(x) = 1, \qquad \lim_{x \to -\infty} f(x) = -1.

Dominant terms

Definition Let p(x)=anxn+an−1xn−1+⋯+a1x+a0p(x) = a_n x^n + a_{n - 1} x^{n - 1} + \dots + a_1 x + a_0 with n≥1n \ge 1 and an≠0a_n \neq 0. For x≠0x \neq 0,

p(x)=anxn(1+an−1anx+⋯+a0anxn),p(x) = a_n x^n \left(1 + \frac{a_{n - 1}}{a_n x} + \dots + \frac{a_0}{a_n x^n}\right),

and the bracket tends to 11 as x→±∞x \to \pm\infty. Hence

lim⁡x→±∞p(x)anxn=1.\lim_{x \to \pm\infty} \frac{p(x)}{a_n x^n} = 1.

The term anxna_n x^n is called the dominant (or leading) term.

In plain words Far out, the highest power is all that matters. The other terms still grow, but they become a vanishing fraction of the total.

Example Here is how p(x)=2x3−100x2p(x) = 2x^3 - 100x^2 compares with its leading term.

xxp(x)p(x)p(x)2x3\frac{p(x)}{2x^3}
100010001.9×1091.9 \times 10^90.950.95
10000100001.99×10121.99 \times 10^{12}0.9950.995

The −100x2-100x^2 term is huge in absolute terms, but its share keeps shrinking.

Polynomial and rational growth

Definition Let r=p/qr = p/q, where pp has degree mm and leading coefficient a≠0a \neq 0, and qq has degree nn and leading coefficient b≠0b \neq 0. As x→∞x \to \infty and as x→−∞x \to -\infty,

  1. if m<nm < n, then r(x)→0r(x) \to 0,
  2. if m=nm = n, then r(x)→a/br(x) \to a/b,
  3. if m>nm > n, then r(x)r(x) has no finite limit, because ∣r(x)∣\lvert r(x) \rvert grows without bound.

To see this, divide the numerator and the denominator by xnx^n.

In plain words Compare the top degree with the bottom degree. A bigger bottom wins and sends the quotient to 00. Equal degrees give the ratio of the leading coefficients.

Example Dividing top and bottom by x2x^2,

lim⁡x→∞3x2−xx2+4=lim⁡x→∞3−1x1+4x2=3,\lim_{x \to \infty} \frac{3x^2 - x}{x^2 + 4} = \lim_{x \to \infty} \frac{3 - \frac{1}{x}}{1 + \frac{4}{x^2}} = 3,

while the bottom has the higher degree in

lim⁡x→∞5x+2x2−1=0.\lim_{x \to \infty} \frac{5x + 2}{x^2 - 1} = 0.

The limit of (1 + 1/x) to the power x and the number e

Definition The limit

lim⁡x→∞(1+1x)x=e≈2.718281828\lim_{x \to \infty} \left(1 + \frac{1}{x}\right)^x = e \approx 2.718281828

exists, and the same limit holds as x→−∞x \to -\infty. Here is why, starting from the fact that ln⁡(1+t)/t→1\ln(1 + t)/t \to 1 as t→0t \to 0, from the lesson on special limits. For x>0x > 0 put t=1/xt = 1/x. As x→∞x \to \infty, t→0t \to 0 through positive values, so

xln⁡(1+1x)=ln⁡(1+t)t→1.x \ln\left(1 + \frac{1}{x}\right) = \frac{\ln(1 + t)}{t} \to 1.

Now

(1+1x)x=exln⁡(1+1/x),\left(1 + \frac{1}{x}\right)^x = e^{x \ln(1 + 1/x)},

and the exponent tends to 11. To conclude that the power tends to e1=ee^1 = e we need one more fact, that eu→e1e^u \to e^1 whenever u→1u \to 1. This is the continuity of the exponential function at 11, which the continuity unit proves. As x→−∞x \to -\infty the same steps work with x<−1x < -1, where 1+1/x>01 + 1/x > 0 and t→0t \to 0 through negative values. Some texts instead define ee as this limit and prove its existence another way. Then the logarithm limit is derived from it, and the argument above runs in the opposite direction.

In plain words Two effects compete. The base gets closer to 11, which pulls the power toward 11, while the exponent grows, which pushes it up. They balance at ee.

Example The values climb toward ee slowly.

xx1110101001001000100010610^6
(1+1x)x\left(1 + \frac{1}{x}\right)^x222.593742.593742.704812.704812.716922.716922.718282.71828

The same argument gives

lim⁡x→∞(1+2x)x=e2,\lim_{x \to \infty} \left(1 + \frac{2}{x}\right)^x = e^2,

since with t=2/xt = 2/x the exponent is

xln⁡(1+2x)=2⋅ln⁡(1+t)t→2.x \ln\left(1 + \frac{2}{x}\right) = 2 \cdot \frac{\ln(1 + t)}{t} \to 2.

Check yourself

  1. Find the limit
lim⁡x→∞4x3+x2x3−7.\lim_{x \to \infty} \frac{4x^3 + x}{2x^3 - 7}.
Show the answer

The degrees are equal, so the limit is the ratio of the leading coefficients, 4/2=24/2 = 2.

  1. Find the limit
lim⁡x→−∞4x2+1x+3.\lim_{x \to -\infty} \frac{\sqrt{4x^2 + 1}}{x + 3}.
Show the answer

For x<0x < 0 we have ∣x∣=−x\lvert x \rvert = -x, so 4x2+1=−x4+1/x2\sqrt{4x^2 + 1} = -x\sqrt{4 + 1/x^2}. Dividing top and bottom by xx gives

4x2+1x+3=−4+1/x21+3/x,x<0, x≠−3,\frac{\sqrt{4x^2 + 1}}{x + 3} = \frac{-\sqrt{4 + 1/x^2}}{1 + 3/x}, \qquad x < 0, \ x \neq -3,

which tends to −2-2.

  1. The limit of 5/x5/x as x→∞x \to \infty is 00. For ε=0.01\varepsilon = 0.01, find an MM that works.
Show the answer

For x>0x > 0 we need 5/x<0.015/x < 0.01, that is x>500x > 500. So M=500M = 500 works, as does any larger number.

How this lesson connects

It also uses rational functions.