This action might not be possible to undo. Are you sure you want to continue?

n 1.1. Deﬁnition: a series of functions of the form n=1 an x where a1 , a2 , . . . , are given (complex) numbers, is called a power series (in x). Partial sums: Sn (x) = a0 + a1 x + · · · + an xn is a polynomial of degree (at most) n. ∞

Example. For S5 (x) = x + 2x4 .

∞ n=1

nxn , the coeﬃcient an is given by an =

2

0, √

n,

n is not a square . In this example n is a square

1.2. Deﬁnition: the set of real (or complex) values of x for which the power series converges is called the domain of convergence. We’ll denote it by D: D = {x ∈ R : n an xn convergent}. Note: 0 ∈ D, so D is never empty. 2. Radius of convergence 2.1. Deﬁnition: the radius of convergence R is the positive number (or +∞) given by 1 = lim sup |an |1/n R n 2.2. Theorem [Cauchy-Hadamard]. The power series n=1 an xn : a) converges absolutely for |x| < R. b) diverges for |x| > R c) Moreover, the convergence is uniform on |x| ≤ R − δ, for any δ > 0. Note: the case of |x| = R is not covered by the theorem, and needs to be analyzed separately. Note: if R = +∞ then D = R. 3. How to compute R. In most of the speciﬁc cases |an |

1/n ∞

is actually convergent, in which case lim sup = lim.

**3.1. First method: keep in mind the following facts:
**

1 • (Average) xn → L ⇒ n (x1 + x2 + · · · + xn ) → L • (Multiplicative analogue) xn > 0, xn → L ⇒ limn (x1 x2 . . . xn )1/n = L. 1/n • Corollary: if xn > 0 and xn+1 /xn → L, then xn → L.

**3.2. Second method: let bn = log(|an |1/n ) =
**

∞

1 n

log |an |. Check if bn has a limit, then |an |1/n → elim bn .

n n n 3.3. Example. n=1 (−1) n2 x . Coeﬃcients: an = (−1)n n2n ; |an |1/n = 2n1/n . What is lim supn→+∞ 2n1/n ? Take log: bn = log |an |1/n = ln 2 + ln n → ln 2. Then |an |1/n → elog 2 = 2. n Radius: 1/R = lim |an |1/n = 2 ⇒ R = 1/2. Hence (−1/2, 1/2) ⊆ D. Check ”endpoints” separately: At x = −1/2: an xn = n divergent. At x = 1/2: an xn = (−1)n n divergent. Therefore D = (−1/2, 1/2). x 3.4. Example. n=1 n! . Coeﬃcients: an = 1/n!, |an |1/n = 1/(n!)1/n . Take log: bn = log |an |1/n = − log n! . Not so clear what lim bn is. n 1 Alternative: |an+1 |/|an | = 1/(n+1)! = n+1 → 0, hence |an |1/n → 0 as well. 1/n! Conclusion: R = +∞, D = R. ∞

n

3.5. Example. but lim sup |an |1/n

n1/2n , n is a square . In this case |an |1/n is not convergent, 0, otherwise = 1. (Justify this step.) So R = 1. It is not hard to see that D = (−1, 1). nxn . Here |an |1/n =

2

1

2. Then: a) f (x) is diﬀerentiable on (−R. Let f (x) = an xn for x ∈ D. ∞ n=1 nan xn−1 . r] ⊂ (−R. R) .2 4. Since n1/n → 1. (The limit f (R−) exists and equals the sum of the series at x = R. for |x| < 1 (R = 1). If R ∈ D. R). for x ∈ R. Then clearly f (x) = x 1 dt 0 1+t2 1 1+x2 .) Note: the theorem is works just the same for −R instead of R. we have Theorem. The only function that satisﬁes this equation is f (x) = ex . n=1 n 5. Assume R is the radius of convergence of the power series an xn . Example. for |x| < 1. Note: applying this theorem repeatedly we see that the sum of a power series is a smooth (inﬁnitely many times diﬀerentiable) function inside the radius of convergence. and f (x) = is f (x) = f (x).1. R). n=0 n! = e. x n n−1 1 4. x2n+1 = = tan−1 (x). Since the power series converges uniformly on compact intervals of the form [−r. R) and f (x) = x ∞ an b) 0 f (t)dt = n=0 n+1 xn+1 . for |x| < 1. Then f is C on R. Note that if one could take x = 1 in the previous two examples.4. then limx→R− f (x) = n=0 an Rn . Example. and let f (x) the sum of the power series ∞ inside the radius of convergence. Let R the radius of convergence of n=0 an xn . 4. ln(1 + x) = ∞ (−1)n+1 xn . Example. n ∞ 4.1. ∀x ∈ (−R. Moreover f (0) = 1. How to use power series 4. Let f (x) = . n n=1 This step is justiﬁed by the following theorem: Theorem[Abel]. In ∞ 1 particular. Let f (x) = Integrate: ∞ (−1) n=0 2n+1 n (−1)n x2n . ∀x ∈ (−R. ∞ ∞ π (−1)n = 2n + 1 4 n=0 ∞ . it is not hard ∞ ∞ an n+1 n−1 to see that the two series n=1 nan x and n=0 n+1 x have the same radius of convergence R. that n=1 n! x n! . R). we would obtain nice identities: (−1)n+1 = ln 2.3. Abel’s theorem 5. f ∈ C ∞ (−R.

- 2014 Standard 1 - Water Supply
- Answer Key 2
- Set
- Engineering Boards
- ME Chapter 1
- ME Chapter 2
- Ice - Chapter 3
- Ice - Chapter 4
- Ice - Chapter 2
- Ice - Chapter 4
- Certificate Para Sa HSS
- 4 Reflection
- 156
- 1
- Shanelyn Cur
- D300102-M
- Glossary of Hydro Power Terms
- Drying
- 5 e NBI Clearance
- 560AA
- Faculty of Engineering School of Mechanical And
- 976AA
- 000329
- 106JA
- Water Calculation Worksheet

Sign up to vote on this title

UsefulNot usefulfdfdf fdfdgf gfgf

fdfdf fdfdgf gfgf

- 9-2
- 9-3
- 9-1
- Solve O.D.E by Series Expansion
- Partial Differential Equation
- Taylor Series[1]
- Green's theorem
- To Find an Orthonormal Basis
- limit[最大整數函數]
- QFFF Thesis
- QFFF Thesis 09
- A note on q-nanlogue of Sandor's functions
- Further Results On The Basis Of Cauchy’s Proper Bound for the Zeros of Entire Functions of Order Zero
- Zeta Function Regularization
- Regular Variation Talk
- geostb
- Novikov (2004)
- Sergey Astashkin and Konstantin Lykov- Extrapolation description of rearrangement invariant spaces and related problems
- ant13-8.pdf
- The kernels of the Hankel and Toeplitz operator
- Ver Sich
- On Ramanujan's Continued Fraction for (q2; q3)Inf (q;q3)Inf
- SSRN-id139549 Fernh
- On Certain Classess of Multivalent Functions
- The Gamma Function
- Recall
- Norms
- Stochastic Calculus for Discontinuous Processes
- Tute 6 - M2962
- Queueing Model
- Power Series