<<

Math 231 Exam 3 Review Sheet by Ruth Luo1

Concepts Covered: Alternating , , , Taylor/Maclaurin Series, Taylor poly- nomials and error

Alternating Series

P P n P P n+1 • Consider a series an = (−1) bn or an = (−1) bn. If

1. bn > 0 for all n,

2. bn is decreasing (i.e., bn > bn+1) for all n, and

3. limn→∞ bn = 0, P then an converges.

P n P n+1 • If S = (−1) bn or S = (−1) bn is an that satises the three conditions above, then the error RN obtained when approximating S with partial sum SN is upper bounded by bN+1, that is,

|RN | < bN+1.

P P P • A series an converges absolutely if |an| converges (this also imples that an converges). P P P • A series an converges conditionally if |an| diverges but an converges. • Exercises 1. Show that P∞ n n+1 converges conditionally. n=0(−1) n2+1 2. What is the smallest value of such that the partial sum approximates the sum P∞ (−1)n N SN S = n=1 n2 within 1 ? 200 Convergence tests

• The following are the new convergence tests we have learned. Note that it is still important to remember the old convergence tests for series ( comparison test, comparison test, test).

P an+1 • - for a series an, let L = limn→∞ | |. Then an P  if L < 1, then an absolutely converges, P  if L > 1, then an diverges,  if L = 1, then the test is inconclusive. P 1/n • - for a series an , let L = limn→∞ (|an|) . Then the same conditions hold as in ratio test. Ratio test is useful when contains . Note that (n+1)! (n+1)·(n)·(n−1)·...·2·1 . • an n! = (n)·(n−1)·...·2·1 = n + 1

n 1/n • Root test is useful when an contains nth powers. E.g., (|c |) = c. • Exercises - determine if the following series converge absolustely, converge conditionally, or diverge. State explicitly which test you are using.

1. P∞ (−3)nn2 n=1 n! 2. P∞ (−2)2n n=1 nn ∞ 3. P (−1)n √ n n=1 n3+5 4. P∞ n 5nn n=1(−1) n! Power Series [email protected]; there may be typos.

1 A power series is a of the form P∞ n for some coecients . We say that the power series • n=0 cn(x − a) cn is centered at a. There are three options for the convergence of P∞ n: • n=0 cn(x − a)  P∞ n converges for all values of , n=0 cn(x − a) x  P∞ n converges only when , n=0 cn(x − a) x = a  P∞ n has a , i.e., if , then P∞ n converges. n=0 cn(x − a) R |x − a| < R n=0 cn(x − a) If , then P∞ n diverges. |x − a| > R n=0 cn(x − a) • Use ratio test or root test to nd the radius of convergence. When R < ∞, it is necessary to check the convergence on the endpoints, a + R and a − R individually. • The most basic example of a power series is

∞ 1 X = xn, which converges if |x| < 1(R = 1). 1 − x n=0

From this, we can get the power series of many other functions like 1 (by integrating the power • 1+x2 , arctan x series of 1 , 1 (by dierentiating the power series of 1 , (by integrating the power series of 1+x2 (1−x)2 1−x ) ln(1 + x) 1 ), etc. 1+x • Dierentiating and integrating power series do not change the radius of convergence. • Exercises 1. Find the radius of convergence and the of convergence of the power series P∞ 4n(x−1)n . n=0 n+1 Find the power series centered at 0 of the following functions, then state the radius of convergence: 2. 1 1+x2 3. arctan x 4. x3 arctan(4x2) 5. R x3 arctan(4x2)dx 6. x2 (5−x)3 7. x 1+5x3 8. R 1/10 x 0 1+5x3 dx 9. Write down a number which approximates R 1/10 x to within −8. 0 1+5x3 dx 10 Taylor/Maclaurin series

• A is a power series representation of a function centered at some a. Given an innitely dier- entiable function f(x), the general form of the Taylor series of f is

∞ X f (n)(a) f(x) = (x − a)n. n! n=0

(n) A Maclaurin series is a Taylor series where , i.e., P∞ f (0) n • a = 0 f(x) = n=0 n! x . • You should know the series of the following functions:  sin x =

 cos x =

2  ex =

 arctan x =

 () (1 + x)k =

• Exercises Find the Taylor series of the given functions centered at a. 1. f(x) = x3 cos(x2), a = 0 2. f(x) = cos x, a = π/2 3. f(x) = ln x, a = 2 √ 4. f(x) = 4 1 + x, a = 0 5. Let f(x) = x3 cos(x2). Using the power series obtained in (1), nd f (11)(0). 6. We know that the power series centered at of 1 is P∞ n. Show that this is the same a = 0 f(x) = 1−x n=0 x (n) series we get when we use (a) P∞ f (0) n, and (b) binomial series, to nd the power series of . n=0 n! x f(x) 7. Find the sum of π2 π4 π6 1 − 222! + 244! − 266! + ... 8. Find the sum of 52 53 54 −5 + 2! − 3! + 4! − ... 9. Evaluate 1−cos(2x) . limx→0 1+2x−e2x 2 (Hint: write out a couple terms of the numerator and the denominator. Note that a0+a1x+a2x +... limx→0 2 = b0+b1x+b2x +... a0+0+0+... = a0 .) b0+0+0+... b0 10. Evaluate sin(x2)−x2 cos x . limx→0 x4 Taylor and error

(n) Given a Taylor series for P∞ f (a) n, we have that • f(x) = n=0 n! (x − a)

f(x) = TN (x) + Rn(x)

(n) where PN f (a) n is called the degree Taylor (initial chunk of the Taylor TN (x) = n=0 n! (x − a) N series that has degree at most N) and

f (N+1)(z) R (x) = (x − a)N+1 N (N + 1)!

for some z between a and x. In general, we cannot solve for z. Instead, we can nd an upper bound (a constant) for RN (x) for all x in some interval.

∗ ∗ ∗ • Given an interval, say [c, d], to maximize RN (x), we pick a values of z and x in [c, d] such that z maximizes (N+1) ∗ (N+1) and ∗ maximizes N+1. And so f (z ) ∗ N+1 . f (z) x (x − a) |RN (x)| ≤ | (N+1)! (x − a) | • Exercises

x 1. Find the degree 4 Taylor polynomial, T4(x) of e centered at a = 0.

2. Find the degree 4 Taylor polynomial, T4(x) of sin x centered at a = 0.

3. Let f(x) = sin x. Use Taylor's Theorem to nd the maximum error in the f(x) ≈ T4(x) on the interval [0, 1/2] (where a = 0). x 4. Let f(x) = e . Use Taylor's Theorem to nd the maximum error in the approximation f(x) ≈ T999(x) on the interval [0, 1/10] (where a = 0).

3