9.1 Sequences Sequence = a function whose domain is the set of positive #’s Ex. 1 = a1 2= a2 3= a3 4= a4 …. n = an ** a’s represent terms – can be mapped on a graph by plotting points Recursively defined sequence – when you take the previous term and use it to find the next term Ex. a1 = 3 ak+1 = 2(ak -1) LIMITS OF A SEQUENCE Definition: Let L be a real #. The limit of a sequence {an} is L written as: lim ๐๐ = ๐ฟ ๐→∞ ๏ท ๏ท ๏ท If for each ε > 0 there exists a M>0 such that |๐๐ − ๐ฟ| < ε whenever n> M If the limit L of a sequence exists then the sequence CONVERGES to L. If the limit of a sequence does not exist then the sequence DIVERGES Look at figure ** If a sequence {an} agrees with a function f at every positive integer, and if f(x) approaches a limit L as x๏ ∞, the sequence must converge to the same limit L THEOREM: LIMIT OF A SEQUENCE a) Let L be a real #. Let f be a function of a real variable such that lim ๐(๐ฅ) = ๐ฟ ๐ฅ→∞ b) If {an} is a sequence such that f(n) = an for every positive integer n, then lim ๐๐ = ๐ฟ ๐→∞ THEOREM 5.15 A LIMIT INVOLVING e ๐ ๐+๐ ๐ ) ๐→∞ ๐ ๐ฅ๐ข๐ฆ (๐ + ๐)๐ = ๐ฅ๐ข๐ฆ ( ๐→∞ Commonly Occurring Limits: 1. 2. ln ๐ lim ๐→∞ ๐ = 0 ๐ lim √๐ = 1 ๐→∞ 1 3. 4. lim ๐ฅ ๐ = 1 ๐→∞ (x>0) lim ๐ฅ ๐ = 0 (|๐ฅ| < 1) ๐→∞ = e 5. 6. ๐ฅ lim (1 + ๐)๐ = ๐ ๐ฅ ๐→∞ ๐ฅ๐ ๐→∞ ๐! lim (any x) = 0 (any x) **Ways in which a sequence can fail to have a limit: 1) Terms of the sequence increase without bound: ๐ฅ๐ข๐ฆ ๐๐ = ∞ ๐→∞ 2) Terms of the sequence decrease without bound: ๐ฅ๐ข๐ฆ ๐๐ = −∞ ๐→∞ Convergence Vs. Divergence Ex. ๐๐=(−1)๐ ( Ex. ๐๐ = ๐ ) ๐+1 3๐2 −๐+4 2๐2 +1 Using L’Hopital’s rule to determine Convergence ๐2 Ex. Show that the sequence whose nth term is an = 2๐ −1 converges Factorial (n!) – defined as n! = 1·2·3·4·…·(n-1)·n 0! = 1 (2n)! vs 2n! Squeeze Theorem for Sequences If lim ๐๐ = L = lim ๐๐ and there exists an integer N such that an ≤ cn ≤ bn for all n>N then the ๐→∞ ๐→∞ lim ๐๐ = L ๐→∞ Ex. 5 p. 597 in book (−1)๐ ๐! is squeezed between 2 sequences that converge to 0 ๐๐ ๐→∞ ๐! *** For any fixed #k ๐ฅ๐ข๐ฆ = 0 *** -- the factorial function grows faster than any exponential f(x)! Theorem 9.4 Absolute Value Theorem – if the absolute value sequence converges to 0, the original signed sequence also converges to 0: For sequence {an} if lim |๐๐ | = 0 then lim ๐๐ = 0 ๐→∞ 1 1 ๐→∞ 1 Ex. 1, -½, 22 , -23 , … (-1)n2๐ if you take the absolute value of each term, the sequence converges to 0 so: lim [(−1)๐ ๐→+∞ 1 ] 2๐ =0 Pattern Recognition for Sequences – describes the nth term after discovering a pattern Ex. Finding the nth term for a sequence: 8 −26 80 −242 , , 2 6 24 120 -2, , Steps: Look at numerator first – appears to match 3n – 1 Look at denominator next – alternates between +/- so know (-1)n The denominator also resembles a factorial pattern so n! Final pattern matches: 3๐ −1 (−1)๐ (๐!) Now find the limit of the pattern: lim |๐๐ | = lim (−1)๐ ๐→∞ ๐→∞ 3๐ −1 = ๐! 0 ๏ so converges to 0 Monotonic Sequences and Bounded Sequences ๏ท So far we have determined the convergence of a sequence by finding its limit ๏ now we will provide a test for convergence of sequences without determining the limit Definition of a Monotonic Sequence – a sequence {an} is MONOTONIC if its terms are nondecreasing (each successive term is larger) a1≤ a2≤ a3≤ …≤an or its terms are nonincreasing (each successive term is smaller) Look at Graphs: **Another way to see that the sequence is monotonic is: ๏ to argue that the derivative of the corresponding function is positive for all x ๏ this implies that f is increasing which implies that {an} is increasing. Definition of a Bounded Sequence 1. A Sequence {an} is bounded above if there is a real #M such that an ≤ M for all n. The number M is called the UPPER BOUND of the sequence. 2. A Sequence {an} is bounded below if there is a real #N such that N ≤an for all n. The number N is called a LOWER BOUND of the sequence. 3. A sequence {an} is bounded if it is bounded above and bounded below. Complete Property of Real Numbers: means there are no holes or gaps on the real # line. ๏ท Used to conclude that if a sequence has an upper bound, it must have a LEAST UPPER BOUND (an upper bound that is smaller than all other upper bounds for the sequence) THEOREM 9.5 Bounded Monotonic Sequences ๏ท If a sequence {an} is bounded and monotonic ๏ then it CONVERGES Ex. an = bn = 1 ๐ is bounded and monotonic – so it converges ๐2 ๐+1 is monotonic but not bounded (only below) cn = (-1)n is bounded by not monotonic Ex. Determine if the sequence with the given nth term is monotonic. Discuss the boundedness 1 #83. an= 4 - ๐ a1 4-1 = 3 a3 4 - โ = 3.66 a2 4-½ = 3.5 a4 4 - ¼ = 3.75 ** as you look at each subsequent term, it is getting larger ๏ SO MONOTONIC ** also see if graph – that it becomes bounded above and below 9.2 Series and Convergence INFINITE SERIES: ∑∞ ๐=๐ ๐๐ = a1 + a2 + a3 + …+ an Definitions of Convergent and Divergent Series ๏ท ๏ท ๏ท ๏ท For the infinite series ∑∞ ๐=1 ๐๐ , the nth partial sum is given by: Sn = a1 + a2 +…+an If the sequence of partial sums {sn} converges to S, then the series above CONVERGES The limit S is called the sum of the series If {sn} diverges then the series DIVERGES Ex. Find the first 5 terms of the sequence of partial sums: 1 + ¼ + 1/9 + 1/16 + 1/25 +…+ S1 = 1 S2 = 1 + ¼ = 1.25 S3 = 1 + ¼ + 1/9 = 1.3611 S4 = 1 + ¼ + 1/9 + 1/16 = 1.4236 S5= 1 + ¼ + 1/9 + 1/16 + 1/25 = 1.4636 Ex. Convergent Series: ∑∞ ๐=1 1 2๐ = ½ + ¼ + 1/8 +1/16 + … 2๐ −1 ๐→∞ 2๐ So lim s1 = ½ s2 = ½ + ¼ = ¾ s3 = ½ + ¼ + 1/8 = 7/8 sn= = 1 so Converges Ex. ∑∞ ๐=1 1 = 1 + 1 + 1 +1 ๏ sn = n which diverges 2๐ −1 2๐ Telescoping Series: (b1 – b2) + (b2 - b3)+ (b3 – b4) +… **sn = b1- bn+1 ** a telescoping series will converge if and only if bn approaches a finite number as n๏ ∞ ** Also, if the series converges, its sum is S = b1 - lim ๐๐+1 ๐→∞ *** If series converges then: Sum = b1 - ๐ฅ๐ข๐ฆ ๐๐+๐ ๐→∞ 2 Ex. Find the sum of ∑∞ ๐=1 4๐2 −1 2 Step 1: Use partial fractions and rewrite as: an = 4๐2 −1 ๏ 2 = A(2n+1) - B(2n – 1) 1 1 So an = 2๐−1 - 2๐+1 1 Sn = 2(1)−1 − ∑∞ ๐=1 2 4๐2 −1 1 2(1)+ 1 + 1 1 − 2(2)+ 1 + 2(2)− 1 1 = lim ( 1 − 2๐+1) = 1 ๐→∞ 1 … + = 1 - 2๐+1 so Converges ๐ 2 n Geometric Series -- ∑∞ a≠0 is a geometric series with ratio r ๐=๐ ๐๐ = a + ar + ar +…+ ar Theorem: Convergence of a Geometric Series *a geometric series with ratio r diverges is |๐|≥ 1. If 0 < |๐|< 1 then the series converges to the sum ๐ ∑∞ ๐=๐ ๐๐ = ๐ ๐−๐ 0<|๐|<1 Ex. Verify if it diverges or converges a) 3 n ∑∞ ๐=0 3(2) ๐! b) ∑∞ ๐=1 2๐ ** Make sure that you know the difference between an infinite series and a sequence: ๏ท ๏ท Infinite Series = an infinite sum of terms form a sequence a1+a2+…+an Sequence = an ordered collection of numbers a1,a2,a3…an Theorem 9.7 Properties of Infinite Series ๏ท If ∑ ๐๐ = A and ∑ ๐๐ = B and c is areal number, then the following series CONVERGE to the indicated series a) ∑∞ ๐=1 ๐๐๐ = cA b) ∑∞ ๐=1(๐๐ + ๐๐ ) = A+B ∞ c) ∑๐=1(๐๐ − ๐๐ ) = A-B Theorem 9.8 Limit of nth Term of a convergent series ๏ท ∑∞ ๐=1 ๐๐ converges, then lim ๐๐ = 0 ๐ผ๐ **converse is generally not true ๐→∞ Theorem 9.9 nth term test for Divergence ๏ states that it the limit of the nth term of a series does not converge to 0, the series must diverge. *If ๐ฅ๐ข๐ฆ ๐๐ ≠0 then ∑∞ ๐=๐ ๐๐ diverges ๐→∞ ๐ Ex. ∑∞ ๐=0 2 ๐! Ex. ∑∞ ๐=1 2๐!+1 1 Ex. ∑∞ ๐=1 ๐ 9.3 The Integral Test ๏ applies only to series with positive terms Theorem 9.10 The Integral Test ๏ท ∞ If f is positive, continuous, and decreasing for x≥1 and an = f(n) then ∑∞ ๐=1 ๐๐ and ∫1 ๐(๐ฅ)๐๐ฅ either both converge or both diverge 2 Ex. ∑∞ ๐=1 3๐+5 2 Let f(x) = 3๐+5 x=1 2 3(1)+ 5 2 =8 x=2 2 3((2)+ 5 ๏ +/Contin. And decreasing ∞ Integrate so: ∫1 #8. ln 2 √2 + ln 3 √3 ∑∞ ๐=2 ln ๐ √๐ + ln 4 √4 + ln 5 √5 2 3๐ฅ+5 2 ๐๐ฅ = 3 [ln(3๐ฅ + 5)]∞1 = ∞ Diverges +… ln ๐ฅ √๐ฅ so f(x) = Step #1. Derive f´(x)= 2−ln ๐ฅ 3 ๏ f is positive and continuous, and decreasing for x>2 2๐ฅ 2 ∞ ln ๐ Step #2. Integrate ∫2 #11 ∑∞ ๐=1 √๐ฅ dx = [2√๐ฅ (ln ๐ฅ − 2)]∞2 = ∞ DIVERGES 1 √๐+1 P-Series and Harmonic Series - has a simple arithmetic test for convergence or divergence ๐ P-Series ∑∞ ๐=๐ ๐๐ = ๐ ๐๐ ๐ ๐ + ๐๐+ ๐๐ + … is a p-series where p is a positive constant 1 For p = 1 ∑∞ ๐=1 ๐ = 1+ ½ + 1/3 +…+ = HARMONIC SERIES ๏ท 1 General Harmonic Series is of the form ∑ (๐๐+๐) 1 1 1 1 Theorem 9.11 Convergence of P-Series: ๏ The p-series ∑∞ ๐=1 ๐๐ = 1๐ + 2๐ + 3๐ + … ๏ท ๏ท Converges if p>1 Diverges if 0 < p ≤1 Ex. discuss the convergence/divergence of: a) Harmonic series ∑∞ ๐=1 b) P series with p = 2 1 ๐ = 1 + ½ + 1/3 +… 1 ∑∞ ๐=1 12 + 1 1 + 22 32 +… Testing for Convergence ∑∞ ๐=2 1 ๐ ln ๐ ๏ 1 ๐ฅ ln ๐ฅ is positive and continuous for x≥2 DIVERGES b/c ≤ 1 p>1 so CONVERGES 1+ln ๐ฅ Steps: 1. find derivative (๐ฅ ln ๐ฅ)−1 ๏ f´(x) = (-1)(๐ฅ ln ๐ฅ) −2(1+ln x) = − ๐ฅ 2 (ln ๐ฅ 2 ) 2. f´(x)<0 for x>2 so now do integral test ∞ 1 ∫2 ๐ฅ ln ๐ฅ 1 ๐๐ฅ = ∞ ∫2 ln๐ฅ๐ฅ dx = lim [ln(ln ๐ฅ)]b2 ๏ lim [ln(๐๐๐) − ln(๐๐2] = ∞ ๐→∞ ๐→∞ 9.4 Comparisons of Series ๏ positive term series ๏ท Past convergence tests a) Terms of series have to be fairly simple b) Series must have special characteristics in order for convergence tests to be applied ๏จ A slight deviation from these special characteristics can make a test nonapplicable **the next tests allow you to compare a series having complicated terms with a simple series whose convergence or divergence is known. Theorem 9.12 Direct Comparison Test Let 0 < an ≤bn for all n ∞ 1. If ∑∞ ๐=๐ ๐๐ converges, then ∑๐=๐ ๐๐ converges ∞ 2. If ∑∞ ๐=๐ ๐๐ diverges, then ∑๐=๐ ๐๐ diverges 1 Ex. ∑∞ ๐=1 3๐2 +2 Ex. ∑∞ ๐=1 1 3 4 √๐−1 1 1 (this is similar to 3๐2 ) ๏ like 1/3∑∞ ๐=1 ๐2 which is a p series where p>1 so Converge 1 compare with p series: 1/4∑∞ ๐=1 4 √๐ diverges b/c p<1 (1/4) Limit Comparison Test – a given series closely resembles a p-series or a geometric series, but you cannot establish the term by term comparison necessary to apply the Direct Comparison Test ๏ so you use now a LIMIT COMPARISON TEST Theorem 9.13 LIMIT COMPARISON TEST ๐ *Suppose that an>0, bn>0 and lim (๐๐ ) = L ๐→∞ ๐ L is finite and positive. Then the two series ∑ ๐๐ and ∑ ๐๐ either both converge or both diverge. 1 Ex. Show that ∑∞ ๐=1 ๐๐+๐ a>0 and b>0 ๏ท By direct comparison ∑∞ ๐=1 1 ๐ (a divergent harmonic series) ๏ท You have lim ๐→∞ 1 ๐๐+๐ 1 ๐ ๐ ๐๐+๐ ๐→∞ = lim 1 =๐ ๏ because the limit is greater than 0, you can conclude the given series diverges A LIMIT COMPARISON TEST WORKS WELL FOR COMPARING A “MESSY” ALGEBRAIC SERIES WITH A PSERIES! *In choosing an appropriate p-series, you must choose one with an nth term of the same magnitude as the nth term of the given series. Given Series Comparison Series Conclusion ∑∞ ๐=1 1 3๐2 −4๐+5 ∑∞ ๐=1 1 ๐2 Both Series Converge ∑∞ ๐=1 1 √3๐−2 ∑∞ ๐=1 1 √๐ Both Series Diverge ∑∞ ๐=1 ๐2 −10 4๐5 + ๐3 ∑∞ ๐=1 ๐2 ๐5 1 = ∑∞ ๐=1 ๐3 Both series converge *** When choosing a series for comparison, you can disregard all but the highest powers of n in both the numerator and denominator *** Ex. Determine the convergence or divergence of: √๐ ๐2 + 1 a) ∑∞ ๐=1 b) ∑∞ ๐=1 c) ∑∞ ๐=1 ๐2๐ 4๐3 +1 1 ๐√๐2 +1 9.5 Alternating Series ๏ท ๏ท Series that contain both positive and negative terms Alternating series occur in 2 ways: the odd terms are negative or the even terms are negative Theorem 9.14 Alternating Series Test: Let an>0. The Alternating Series following 2 conditions are met: 1) lim ๐๐ = 0 ๐→∞ and ∞ ๐ ๐+1 ∑∞ ๐๐ Converge if the ๐=1(−1) ๐๐ and ∑๐=1(−1) 2) ๐๐+1 ≤ an for all n 1 ๐+1 Ex. Determine the convergence or divergence of ∑∞ ๐=1(−1) ๐ a) lim ๐๐ ๏ ๐→∞ b) An+1 = 1 ๐+1 1 ๐ ๐→∞ 1 ๐ lim ≤ = converges to 0 Ex. determine the convergence/divergence of ∞ ∑ ๐=1 ∞ ∑ ๐=1 (−1)๐+1 2๐ − 1 (−1)๐+1 ln(๐ + 1) ๐+1 p. 632 are examples where alternating series test fails ALTERNATING SERIES REMAINDER – the partial sum Sn can be a useful approximation for sum S of the series. The error involved in using S≈Sn is the remainder Rn = S - Sn ๏ท If a convergent alternating series satisfies the condition an+1≤an, then the absolute value of the remainder Rn involved in approximating the sum S by Sn is less than or equal to the first |๐ − ๐๐ | = |๐ ๐ | ≤ an+1 neglected term Ex. ∑∞ ๐=1 (−1)๐+1 3 ๐2 approximate the sum of the series by using the 1st 6 terms Absolute and Conditional Convergence **Sometimes a series may have both + and – terms and not be an alternating series ๏ so to obtain information about the convergence – you must investigate (by direct comparison, etc…) Theorem 9.16 Absolute Convergene ๏ท ๏ท If the series ∑|๐๐ | converges, then the series ∑ ๐๐ also converges. Converse of theorem 9.16 is not true Ex. Alternating harmonic series ∑∞ ๐=1 (−1)๐+1 ๐ 1 2 1 3 1 4 = 1- + - +… Definition of Absolute and Conditional Convergence 1. ∑ ๐๐ is absolutely convergent if ∑|๐๐ | converges 2. ∑ ๐๐ is conditionally convergent if ∑ ๐๐ converges but ∑|๐๐ | diverges Ex. Determine whether the series converges conditionally or absolutely diverges #49. ∑∞ ๐=1 So (−1)๐+1 √๐ = −1 √๐+1 1 √๐ < However ∑∞ ๐=1 (−1)1+1 √1 + (−1)2+1 … √2 1 ๐→∞ √๐ and lim 1 √๐ =0 1 1 √2 =1 - so by alternating series test CONVERGES (p=1/2 which is p<1 ๏ so diverges) ***SO SERIES CONVERGES CONDITIONALLY *** #58. ∑∞ ๐=0 (−1)๐ √๐+4 Rearrangement of Series ๏ in infinite series the value of the sum may change depending on whether the series is absolutely convergent or conditionally convergent. 1 1 ๐+1 Ex. the alternating harmonic series converges to ln 2 ∑∞ =1–½+3-4+… ๐=1(−1) **however, rearrangement of the series produces a different sum: 9.6 The Ratio and Root Tests Test for absolute convergence: Let ∑ ๐๐ be a series with nonzero terms RATIO TEST: 1. 2. 3. ๐๐+1 |<1 ๐๐ ๐ ๐ ∑ ๐๐ diverges if lim | ๐+1 | >1 or lim | ๐+1 | = ∞ ๐ ๐→∞ ๐→∞ ๐๐ ๐ ๐ The ratio test is inconclusive if lim | ๐๐+1 | = 1 ๐→∞ ๐ ∑ ๐๐ converges ABSOLUTELY if lim | ๐→∞ **Ratio Test ๏ is particularly useful for series that converge rapidly (series involving factorials/expoentials) Ex. ∑∞ ๐=0 2๐ ๐! ๐! ๐! 1 = = (๐+1)! (๐+1)๐! ๐+1 2๐+1 2๐ lim [(๐+1)! ÷ ๐! ] ๐→∞ ๏ท Simplifying quotients of factorials: ๏ท An = 2๐ ๐! write as lim | ๐→∞ ๐๐+1 | ๐๐ = 2๐+1 ๐! = lim [(๐+1)! โ 2๐ ] ๐→∞ = lim 2 ๐→∞ ๐+1 Ex. a) ∑∞ ๐=0 ๐2 2๐+1 3๐ = 0 Coverges ๐๐ c) ∑∞ ๐=1 ๐! THE ROOT TEST – works well for series involving nth powers Theorem 9.18 Root test: Let ∑ ๐๐ be a series: ๐ 1. ∑ ๐๐ converges absolutely if ๐ฅ๐ข๐ฆ √|๐๐ | <1 ๐→∞ ๐ ๐ 2. ∑ ๐๐ diverges if ๐ฅ๐ข๐ฆ √|๐๐ | >1 or ๐ฅ๐ข๐ฆ √|๐๐ | = ∞ ๐→∞ ๐→∞ ๐ 3. The root test is inconclusive if ๐ฅ๐ข๐ฆ √|๐๐ | = 1 ๐→∞ **The root test is always inconclusive for any p-series ** Ex. Determine the convergence or divergence of: ∑∞ ๐=1 ๐ ๐ 2๐ ๐๐ ๐ 2๐ ๐ lim √|๐๐ | = lim √ ๐๐ ๐→∞ ๐→∞ 2๐ = lim ๐→∞ ๐๐ ๐ ๐๐ ๐2 ๐→∞ ๐ = lim = 0 which is < 1 ๏ this limit will always be <1 so it Converges absolutely. Guidelines for Testing a Series for Convergence or Divergence: 1. Does the nth term approach 0? If not ๏ the series DIVERGES 2. Is the series one of the special types ๏ geometric, p-series, telescoping, or alternating 3. Can the Integral, Root or Ratio Test be applied 4. Can the series be compared favorably to one of the special types. Look at example #5 p. 643 and make Chart! 9.7 Taylor Polynomials and Approximations p(x) = a0 + a1x + a2x2 + a3x3 + … + anxn *local linear approximation of 1st degree polynomial: f(x) ≈ f(x0) + f´(x0)(x – x0) Polynomial Approximations of elementary functions Goal: To use polynomial functions as approximations of other functions Must: #1. 1st begin by choosing a #c in the domain of f at which f and P have the same value where P(c ) = f(c) ๏ the graphs of both will pass through (c, f(c)) ** approximating polynomial is said to be expanded about c or centered at c ** Many polynomial graphs pass through the point (c, f(c)) ** YOUR TASK: to find the polynomial whose graph resembles the graph of f at this pt. #2. To do this: ignore the additional requirement that the slope of the polynomial function be The same as the slope of the graph of f at the point (c,f(c)) ๏ P´(c) = f´(c) Ex. For function f(x) = ex find the first degree polynomial function P1(x) = a0 + a1x whose value and slope agree with the value and slope of f at x=0. ๏ท ๏ท ๏ท ๏ท Because f(x) = ex f´(x) = ex f(0) = e0 = 1 f´(0) = e0 = 1 ๏ The value of f and the slope of f at x = 0 Because P1(x) = a0 + a1x you can use the condition P1(0) = f(0) so a0 = 1 Also because P1´(x) = a1 you can use the condition that P1´(0) = f´(0) so P1´ = 1 So = P1(x) = 1 + x ๏ at (0,1) the graphs are reasonably close – but as you move away the graphs move farther away from each other ๏ so need to improve approximation even more by: ** that the values of the 2nd derivatives of P and f agree when x=0 has the form P(x) ≈ a0 + a1x + a2x2 P(0) = f(0) = 1 p´(0) = f´(0) = 1x p´´(0) = f´´(0) = ½ x² ๏ p´´(x) = 2a2 so p´´(0) = 2(0) = 0 ๏ p2(x) = 1 + x + ½ x2 **if you continued finding the derivatives you would get: ๐ ๐! ๐ ๐! Pn(x) = 1 + x + ½ x2 + x3 + …+ xn ≈ ex Taylor and Maclaurin Polynomials **for expansion about an arbitrary value of c write the polynomial in the form: Pn(x) = a0 + a1(x-c) + a2(x-c)2 + a3(x-c)3 + …+ an(x-c)n Pn´(x) = a1 + 2a2(x-c) + 3a3(x-c)2 + … + nan(x-c)n-1 Pn´´(x) = 2a2 + 2(3a3)(x-c) + … + n(n-1)an(x-c)n-2 Pn´´´(x)= 2(3a3) + …+ n(n-1)(n-2)an(x-c)n-3 . . . Pnn(x) = n(n-1)(n-2) . . . 2(1)an So letting x=c you get… Pn(c) = a0 Pn´(c) = a1 Pn´´9c) = 2a2 pnn(c) = n!an **Since the value of f and its first n derivatives must agree with the value of P n and its first n derivatives at x=c follows that: f(c) = a0 f´(c) = a1 ๐´´ (๐) 2! = a2 ๐๐ (๐) ๐! = ๐๐ Taylor Polynomial for f at c – if f has n derivatives at c, then the polynomial: Pn(x) = f(c) + f´(c)(x-c) + ๐´´ (๐) 2! (x-c)2 +. . . + ๐(๐) ๐ ๐! (x-c)n Maclaurin Polynomial for f if c = 0 Pn(x) = f(0) + f´(0)x + ๐´´ (0) 2 x 2! + ๐´´´ (0) 3 x 3! +...+ ๐(๐) (0) n x ๐! Ex. Find the Taylor polynomials P0, P1, P2, P3, P4 for f(x) =lnx centered at c=1 F(x) = ln x f(1) = ln 1 = 0 1 1 f´(x) = ๐ฅ f´(1) = 1 = 1 1 f´´(x) = − ๐ฅ 2 2! f´´´(x) = ๐ฅ 3 3! f(4)(x) = -๐ฅ 4 f´´(1) = −1 12 = -1 2! f´´´(1) = 13 = 2 f(4)(1) = −3! = 14 -6 Taylor Polynomials are as follows: P0(x) = f(1) = 0 P1(x) = f(1) + f´(1)(x-1) = (x-1) P2(x) = f(1) + f´(1)(x-1) + ๐´´ (1) (x-1)2 2! = (x-1) – ½ (x-1)2 P3(x) = f(1) + f´(1)(x-1) + ๐´´ (1) (x-1)2 2! + ๐´´´ (1) (x-1)3 3! = (x-1) – ½ (x-1)2 + 1/3(x-1)3 P4(x) = f(1) + f´(1)(x-1) + ๐´´ (1) (x-1)2 2! + ๐´´´ (1) (x-1)3 3! + ๐4 (1) (x-1)4 4! = (x-1) – ½ (x-1)2 + 1/3(x-1)3 – ¼(x-1)4 Find the Maclaurin polynomial P0, P2, P4, and P6 for f(x) = cos x use P6(x) to approximate the value of cos(0.1) F(x) = cos x f(0) = cos 0 =1 f´(x) = - sinx f´(0) = -sin0 = 0 f´´(x) = -cosx f´´(0)= -cos0 = -1 f´´´(x) = sinx f´´´(0) = sin0 = 0 P0(x) = 1 P2(x) = 1 + 0 + ๐´´ (0) 2 x 2! ๏ 1 – ½!x2 P4(x) = 1 + 0 + ๐´´ (0) 2 x 2! + ๐´´´ (0) 3 x 3! + ๐4 (0) 4 x 4! ๏ 1 – ½!x2 + ¼!x4 P6(x) 1 – ½!x2 + ¼!x4 – 1/6!x6 **Using P6 to approximate cos(.1) = 1- ½!(.1)2 + ¼!(.1)4 – 1/6!(.1)6 = .995004165 Look at p. 652 Taylor Poly. For sin(x) expanded about π/6 **Taylor and Maclaurin Polynomials can be used to approximate the value of a function at a specific point. Ex. to approximate the value of ln(1.1) ๏ use taylor poly for f(x) = ln x expanded about c=1 Ex. Use a fourth maclaurin poly to approximate the value of ln(1.1) ๏ (1.1 is closer to 1 than 0 you use Mauclaurin poly for function g(x) = ln (1+x) g(x) = ln(1+x) g´(x) = g(0) = ln(1+0) = ln 1 = 0 1 (1+x)-1 1+๐ g´(0) = (1+0)-1 = 1 g´´(x) = -(1+x)-2 g´´(0) = -(1+0)-2 = -1 g´´´(x) = 2(1+x)-3 g´´´(0) = 2(1+0)-3 = 2 g4(x) = -6(1+x)-4 g4(0) = -6(1+0)-4 = 6 P4(x) = g(0) + g´(0) + ๐´´ (0) 2 x 2! + ๐´´´ (0) 3 x 3! + ๐4 (0) 4 x 4! = x – ½ x2 + 1/3x3 – ¼ x4 so ln(1.1) = ln(1+.1)≈P4(.1) = .0953083 Remainder of a Taylor Polynomial **To measure the accuracy of approximating a function value f(x) by the Taylor Polynomial P n(x) you can use the concept of a remainder Rn(x): f(x) = Pn(x) + Rn(x) remainder exact value approximate value so Rn(x) = f(x) – Pn(x) ๏ the absolute value of Rn(x) is called ERROR associated with approximation ERROR = |๐ ๐ (๐ฅ)| = |๐(๐ฅ) − ๐๐ (๐ฅ)| TAYLOR THEOREM: if a function f is differentiable through order n+1 in an interval I containing c, then for each x in I there exists Z between x and c such that: F(x) = f(c)+ f´(c)(x-c) + ๐´´ (๐) (x-c)2 + 2! ...+ ๐(๐) (๐) (x-c)n ๐! + Rn(x) where Rnx = ๐(๐+1) (๐ง) (x-c)n+1 (๐+1)! **one useful consequence of Taylor’s Theorem is that |๐ ๐ (๐ฅ)|≤ |๐ฅ−๐|๐+1 (๐+1)! max |๐ ๐+1 (๐ง)| where max |๐ ๐+1 (๐ง)| is the maximum value of f(n+1)(z) between x & c Ex. The 3rd Maclaurin polynomial for sinx is given by P3(x) = x - ๐ฅ3 . 3! Use the Taylor Polynomial Theorem to approximate sin(.1) by P3(.1) and determine the accuracy of the approximation. ๐ฅ3 Sin x = x - 3! + R3(x) = x So sin(.1) ≈ .1 – (.1)3 3! ๐ฅ3 3! + ๐4 (๐ง) 4 x 4! where 0 < z < .1 ๏ .1 - .000167 = .099833 Because f4(z) = sin z it follows that the error |๐ 3 (.1)| can be bounded as follows: 0<R3(.1) = sin ๐ง (.1)4 4! < .0001 4! ≈ .000004 So this implies: .099833 < sin (.1) = .099833 + R3(x) < .099833 + .000004 ๏ .099833 <sin.1 < .099837 Ex. Determine the degree of the Taylor Polynomial Pn(x) expanded about c=1 that should be used to approximate ln (1.2) so that the error is less than .001 ๐! F(x) = lnx and the (n+1)st derivative is givne by f(n+1)(x) = (-1)n๐ฅ ๐+1 ๐๐+1 (๐ง) **Using Taylor’s Theorem |๐ ๐ (1.2)| = | (๐+1)! (1.2 − 1)๐+1 | = = ** In this interval, .2๐+1 [๐ง ๐+1 (๐+1)] is < .2๐+1 ๐๐+1 (๐+1) ๐! 1 [ ] ๐ง ๐+1 (๐+1)! .2n+1 where 1 < z < 1.2 .2๐+1 (๐+1) .2๐+1 (๐+1) So you are seeking a value of n such that: < .001 ๏ 1000 < (n+1)5n+1 So the smallest value n that satisfies the inequality is n=3 so you would need the 3 rd Taylor Poly. To achieve desired accuracy. 9.8 Power Series Approximating Functions by Taylor/Maclaurin Polynomials Ex. f(x) = ex ๏ 1st degree ex≈ 1+x 2nd degree ex≈ 1+ x + ๐ฅ2 2! 3rd degree ex≈ 1 + x + ๐ฅ2 2! + ๐ฅ3 3! 4th degree ex ≈ 1 + x+ ๐ฅ2 2! + ๐ฅ3 3! + ๐ฅ2 2! + ๐ฅ3 3! + 5th degree ex ≈ 1 + x + ๐ฅ4 4! ๐ฅ4 4! + ๐ฅ5 5! **the higher degree of the approximating polynomial, the better the approximation becomes **Several important functions, such as f(x) = ex can be represented exactly by an infinite series called ๏ POWER SERIES ๏ฎ In example above, f(x) = ex ๏ for each real # x, it can be shown that the infinite series: ex = 1 + x + ๐ฅ2 2! + ๐ฅ3 3! +…+ ๐ฅ๐ ๐! converges on the right to ex Definition of Power Series – If x is a variable, then an infinite series of the form: ๐ 2 3 n ∑∞ ๐=๐ ๐๐ ๐ = a0 + a1x + a2x + a3x + . . . + anx + … is called a POWER SERIES. ๐ 2 n More generally, an infinite series of the form: ∑∞ ๐=๐ ๐๐ (๐ − ๐) = a0 + a1(x-c) + a2(x-c) +…+an(x-c) + … is called a POWER SERIES CENTERED AT C, where c is a constant. Ex. p. 660 Radius and Interval of Convergence: ๐ *A power series in x can be viewed as a function of x ๏ f(x) = ∑∞ ๐=๐ ๐๐ (๐ − ๐) DOMAIN of f: the set of all x for which the power series converges ๏ท Every power series converges at its center ๏ so c always lies in the domain of f The Domain of a power series can take 3 forms: 1. A single point 2. An interval centered at c 3. An entire real line Theorem 9.20 Convergence of a Power Series: For a power series centered at c, precisely one of the following is true: 1) The series converges only at c 2) There exists a real # R>0 such that the series converges absolutely for |(๐ฅ − ๐)| < R and diverges for |(๐ฅ − ๐)| > R 3) The series converges absolutely for all x *The # R = Radius of Convergence of the power series. a) If the series converges only at c, the radius of convergence is R = 0 b) If the series converges for all x, the radius of convergence is R = ∞ **Interval of Convergence = the set of all values of x for which the power series converges ๐ n Ex. Finding Radius of Convergence of ∑∞ ๐=0 3(๐ฅ − 2) for x≠ 2, let un = 3(x-2) then: ๐ข๐+1 | ๐→∞ ๐ข๐ lim | 3(๐ฅ−2)๐+1 | ๐→∞ 3(๐ฅ−2)๐ = lim | = lim |๐ฅ − 2| = |๐ฅ − 2| by ratio test, the series converges if |๐ฅ − 2| ๐→∞ < 1 and diverges if |๐ฅ − 2| > 1. Therefore Radius of Convergence with series is R =1 Ex. (−1)๐ ๐ฅ 2๐+1 ∑∞ ๐=0 (2๐+1)! ๐ฅ2 (2๐+3)(2๐+2) ๐→∞ = lim Let (−1)๐ ๐ฅ 2๐+1 un= (2๐+1)! ๐ข then lim | ๐+1 | ๐→∞ ๐ข๐ = lim | ๐→∞ (−1)๐+1 ๐ฅ2๐+3 (2๐+3)! (−1)๐ ๐ฅ2๐+1 (2๐+1)! | * for any fixed value of x, this limit is 0. So by the Ratio Test, the series converges for all x, so Radius of Convergence = ∞ ENDPOINT CONVERGENCE: ๏ท 1. 2. 3. 4. 5. 6. Each endpoint must be tested separately for convergence or divergence ๏ as a result, the interval of convergence of a power series can take any 1 of the 6 forms below: Radius: 0 Radius: ∞ Radius: R Radius: R Radius: R Radius: R Finding the Interval of Convergence: ๐ฅ๐ ∑∞ ๐=1 ๐ letting un = ๐ฅ๐ ๐ ๐ข ๏ lim | ๐+1 | ๐→∞ ๐ข๐ = lim | ๐→∞ ๐ฅ๐ (๐+1) ๐ฅ๐ ๐ ๐๐ฅ | = lim |๐+1| = |๐ฅ|<1 ๐→∞ ** so by ratio test, the radius of convergence is R= 1 1) Because the series is centered at 0, it converges in the interval (-1,1) **however, not necessarily the interval of convergence ๏ to determine, you must test for the convergence at each endpoint 1 - When x = 1 you obtain the divergent harmonic series ∑∞ ๐=1 ๐ - When x = -1 you obtain the convergent alternating harmonic series ∑∞ ๐=1 (−1)๐ ๐ - So the interval of convergence for the series is [-1,1) Ex. Find the interval of convergence of ∑∞ ๐=1 ๐ข๐+1 | ๐→∞ ๐ข๐ lim | = lim | ๐→∞ (−1)๐+1 (๐ฅ+1)๐+1 2๐+1 (−1)๐ (๐ฅ+1)๐ 2๐ (−1)๐ (๐ฅ+1)๐ 2๐ 2๐ (๐ฅ+1) | ๐→∞ 2๐+1 | = lim | =| ๐ฅ+1 | 2 un = (−1)๐ (๐ฅ+1)๐ 2๐ ๐ฅ+1 | <1 2 ๏ท By ratio test the sequence converges if | ๏ท R=2 Because the series is centered at x=-1, it will converge in the interval (-3,1) so at the endpoints < 1 or |๐ฅ + 1| < 2, so the radius of convergence is (−1)๐ −2๐ 2๐ ∑∞ = = ∑∞ ๐=0 ๐=0 1 diverges when ๐ 2 2๐ ๐ 2๐ (−1) ๐ ∑∞ = ∑∞ ๐=0 ๐=0(−1) diverges when x = 1 2๐ you have: ∑∞ ๐=0 ๏ท And ๏ท Both of which diverge, so the interval of convergence (-3,1) x = -3 ** HINTS: a) if all n’s cancel out, then x will have a radius b)If n’s don’t cancel – higher degree on top then ∞ ๏ then R = 0 c) if n’s don’t cancel – higher degree on bottom then 0 ๏ then R = ∞ Differentiation and Integration of Power Series Theorem 9.21 Properties of Functions defined by Power Series: ๐ If the function given by f(x) = ∑∞ = a0 + a1(x-c) + a2(x-c)2 + … has a radius of ๐=0 ๐๐ (๐ฅ − ๐) convergence of R > 0, then on the interval (c-R,c+R), f is differentiable (and therefore continuous). Moreover, the derivative and antiderivative of f are as follows: ๐−1 1. f´(x) = ∑∞ = a1 + 2a2(x-c) = 3a3(x-c)2 + … ๐=1 ๐๐๐ (๐ฅ − ๐) 2. ∫ ๐(๐ฅ)๐๐ฅ = C + ∑∞ ๐=0 ๐๐ (๐ฅ−๐)๐+1 ๐+1 (๐ฅ−๐)2 (๐ฅ−๐)3 + a + 2 2 3 = C + a0(x-c) + a1 … Radius of Convergence of Series ๏ obtained by differentiating or integrating is the same as original power series Interval of Convergence ๏ may differ as a result of behavior of endpoints. Ex. Consider the function given by f(x) = ∑∞ ๐=1 ๐ฅ๐ ๐ =x+ ๐ฅ2 2 + ๐ฅ3 3 Find the intervals of convergence for each of the following: a) ∫ ๐(๐ฅ)๐๐ฅ b) f(x) c) f´(x) + …. ๐ฅ ๐+1 ๐ฅ2 ๐ฅ3 ๐ฅ4 And ∫ ๐(๐ฅ)๐๐ฅ = C + ∑∞ ๐=1 ๐(๐+1) = C + 1โ2 + 2โ3 + 3โ4 + … By the ratio test, you can show that each series has a radius of convergence of R=1. Considering the interval (-1,1) you have the following: ๐ฅ ๐+1 a) For ∫ ๐(๐ฅ)๐๐ฅ the series ∑∞ ๐=1 ๐(๐+1) converges for all x = +/-1 and its I of C is [-1,1] ๐ฅ๐ converges for x = ๐ ∞ ๐−1 ∑๐=1 ๐ฅ diverges for x b) For f(x) the series ∑∞ ๐=1 -1 and diverges for x=1 so I of C is [-1,1) c) For f´(x) the series = +/-1 and its I of C is (-1,1) **f´(x) is the least likely to converge at the endpoints ๏จ It can be shown that if the series for f´(x) converges t the endpoints x = c+/- R, the series for f(x) will also converge there 9.9 Geometric Power Series ๐ Remember: Sum of Geometric Series ∑∞ ๐=0 ๐๐ = ๐ 1−๐ |๐| < 1 **In other words if you let a = 1 and r = x, a power series representation for ( 1 1−๐ฅ 1 ) 1−๐ฅ centered at 0 is: ๐ 2 3 = ∑∞ ๐=0 ๐ฅ = 1 + x + x + … |๐ฅ| < 1 1 This series represents f(x) = (1−๐ฅ) only on the interval (-1,1) whereas f is defined for all x≠1 **To represent f in another interval, you must develop a different series ๏ 1 1 1 Ex. to get power series centered at -1 --> 1−๐ฅ = (1−๐ฅ)−1 = 2−(๐ฅ+1)= Which implies a= ½ r = 1 1−๐ฅ 1 ๐ฅ+1 ๐ ) 2 = ∑∞ ๐=0 2 ( ๐ฅ+1 2 ๏ ½ [1 + 1 2 ๐ฅ+1) 1−[ ] 2 so for |๐ฅ + 1| <2 you have: (๐ฅ+1) 2 + (๐ฅ+1)2 4 + (๐ฅ+1)3 8 + …] |๐ฅ + 1| < 2 converges on interval (-3,1) 4 Ex. Find a power series for f(x) = ๐ฅ+2 centered at 0 ๐ a) Write in 1−๐ format = 4 4 2 2 ๐ฅ 2+ 2 = ๐ = 1−๐ 2 1− −๐ฅ 2 ๐ = 1−๐ a = 2 r= ∞ ๐ b) So power series is ๐ฅ+2 = ∑∞ ๐=0 ๐๐ = ∑๐=0 2 −๐ฅ๐ 2 −๐ฅ 2 ๐ฅ = 2( 1 - 2 + ๐ฅ2 4 - ๐ฅ3 + 8 …) −๐ฅ Series converges when | 2 | < 1 and implies interval of convergence = (-2,2) Operations with Power Series: let f(x) = ∑ ๐๐ ๐ฅ ๐ and g(x) = ∑ ๐๐ ๐ฅ ๐ ๐ ๐ 1. f(kx) = ∑∞ ๐=0 ๐๐ ๐ ๐ฅ ๐๐ 2. f(xN) = ∑∞ ๐=0 ๐๐ ๐ฅ ∞ 3. f(x) + g(x) = ∑๐=0(๐๐ ± ๐๐ )๐ฅ ๐ ** All of these operators can change the interval of convergence for the resulting series Ex. for addition ๏ the interval of convergence as the Intersection of the IOC’s of the original series ๐ฅ ∞ ∞ ๐ ๐ ∑∞ ๐=0 ๐ฅ + ∑๐=0(2 ) = ∑๐=0(1 + 1 )๐ฅ ๐ 2๐ (-1,1) Ω (-2,2) = (-1,1) Ex. Adding 2 Power Series: Find a power series centered at 0 for f(x) = 3๐ฅ−1 ๐ฅ 2 −1 3๐ฅ−1 ๐ฅ2− 1 Step 1: Use partial fractions to break down: Step 2: Find each power series 1 ๐ฅ−1 So 3๐ฅ−1 ๐ฅ 2 −1 2 ๐ฅ+1 1 2 1−(−๐ฅ) ๐ |๐ฅ| = 1−๐ฅ = -∑∞ <1 ๐=0 ๐ฅ ๐ ๐ = ∑∞ ๐=0[2(−1) − 1]๐ฅ |๐ฅ| < 1 = and x > -1 2 1 = (๐ฅ+1) + (๐ฅ−1) ๐ ๐ = ∑∞ ๐=0 2(−1) ๐ฅ |๐ฅ| < 1 a = 1 r = -x = 1 – 3x + x2 – 3x3 + x4… IOC is (-1,1) (-1,1) Ex..Finding a power series by integration --Find a power series for f(x) = ln x centered at 1 1 ๐ฅ ๐ ๐ = ∑∞ --๏ a = 1 r = (1-x) ๏ -(x-1) ๐=0(−1) (๐ฅ − 1) 1 [1−(−๐ฅ+1)] 1 ๐ --Integrating produces lnx = ∫ ๐ฅ ๐๐ฅ + ๐ถ = C + ∑∞ ๐=0(−1) (๐ฅ+1)๐+1 ๐ -- letting x =1, then C = 0 so: ln = ∑∞ ๐=0(−1) ( -- Interval of Convergence = (0,2] |−(๐ฅ − 1)|< 1 -x + 1 > -1 ๐+1 ) = ๐ฅ−1 1 (๐ฅ−1)๐+1 ๐+1 – (๐ฅ−1)2 2 + (๐ฅ−1)3 3 – (๐ฅ−1)4 4 +… x>0 x<2 ***Series actually converges at 2 Ex. Find a power series for g(x) = arctanx centered at 0 ๐๐ฆ [arctan ๐ฅ] ๐๐ฅ 1. 1 = (1+ ๐ฅ2 ) so use the following: 1 ๐ ๐ F(x) = 1+๐ฅ ∑∞ ๐=0(−1) ๐ฅ IOC (-1,1) a = 1 r = -x 1 2๐ 2. Substitute x2 for x to get: f(x2) = 1+๐ฅ 2 = ∑∞ -๏ then integrate ๐=0(−1)๐ฅ Get: arctanx = ∫ 1 1+๐ฅ 2 dx + C ๐ = C + ∑∞ ๐=0(−1) ๐ = ∑∞ ๐=0(−1) |−๐ฅ| < 1 ๐ฅ 2๐+1 2๐+1 ๐ฅ 2๐+1 2๐+1 -x > -1 when x =0 C=0 = x- ๐ฅ3 3 + ๐ฅ5 5 - ๐ฅ7 7 IOC = (-1,1) (-1,1) actually converges for x±1 9.10 Taylor Series and MacLaurin Series Theorem 9.22 The Form of a Convergent Power Series ** If f is represented by a power series f(x) = ∑ ๐๐ (๐ฅ − ๐)๐ for all x in an open interval I containing c, then an = ๐(๐) (๐) ๐! and f(x) = f(c) + f’(c)(x-c) + ๐′′ (๐) (x-c)2+ 2! …+ ๐(๐) (๐) ๐! (x-c)n + … **coefficients are the same as Taylor Polynomial so it is called Taylor series for f(x) at c KEY TO THEOREM: says if a power series converges to f(x), the series must be a Taylor Series ๐(๐) (๐) ๐! WHAT THEOREM DOES NOT SAY: that every series formed with the Taylor Coefficients an= will converge to f(x). Definitions of Taylor and MacLaurin Series: If a function f has derivatives of all orders at x=c, then the series: ∑∞ ๐=๐ ๐(๐) (๐) (x-c)n ๐! = f(c) + f’(c)(x-c) + … + ๐(๐) (๐) ๐! (x-c)n +.. Moreover, if c=0, then the series is the MacLaurin Series for f is called the Taylor Series for f(x)at c. **If you know the pattern for the coefficients of the Taylor Polynomials for a function, you can extend the pattern easily to form the corresponding Taylor Series. 1 1 1 Ex. P4(x) of ln x centered at 1 = (x-1) - 2(x-1)2 + 3(x-1)3 - 4(x-1)4 1 2 ๏จ Extend for Taylor Series for ln x centered at c = 1 (x-1) - (x-1)2 +…+ (−1)๐+1 (x-1)n ๐ EX….Use the function f(x) = sin x to form MacLaurin Series and determine the interval of convergence *remember ∑∞ ๐=0 = 0 + 1x + 0x2 - ๐ฅ3 3! ๐(๐) (๐) (x-c)n ๐! + 0x4 + ๐ฅ๐ฅ 5! so power series is ∑∞ ๐=0 + 0x6 - ๐ฅ7 7! = x- ๐ฅ3 3! + ๐ฅ5 5! - ๐ฅ7 7! ๐(๐) (0) n (x) ๐! = f(0) + f’(0)x + ๐′′ (0) 2 x 2! + ๐′′′ (0) 3 x 3! +… +… ๐ฅ 2๐+1 ๐ By Ratio Test: ∑∞ ๐=0(−1) (2๐+1)! converges for all x ** In previous example ๏ CAN’T conclude that the power series converges to sin x for all x ๏จ It converges to some function – but you are not sure what function it is **Important point in dealing with Taylor and MacLaurin Series: 1) You must remember that the derivatives are being evaluated at every single point 2) So another function will agree with the value of f(n)(x) when x=c and disagree at other x-values Figure: **obtain same power series as above ex. ** Series converges for all x, but cannot Converge to both f(x) and sinx for all x --Let f have derivates of all orders in an open interval I centered at c a) Taylor series for f may fail to converge for some x in I b) Or, even if convergent, it may fail to have f(x) as its sum THEOREM 9.23 CONVERGENCE OF TAYLOR SERIES **If lim ๐ ๐ =0 for all x in the interval, I, then the Taylor Series for f converges and equals ∑∞ ๐=0 ๐→∞ c)n ๐(๐) (๐) (x๐! ** Basically this is saying that: a Power Series formed with Taylor Coefficients a n = ๐(๐) (๐) ๐! converges to the function from which it was derived at precisely those values for which the remainder approaches 0 as n๏ ∞ Ex. Show that the MacLaurin Series for f(x) = sin x converges to sin x for all x --From previous example sin x = x - ๐ฅ3 3! + ๐ฅ5 5! - ๐ฅ7 7! ๐ฅ 2๐+1 ๐ + …∑∞ ๐=0(−1) (2๐+1)! is true for all x because f(n+1)(x) = ±sin x or f(n+1)(x) = ±cos x. ๏ฎ You know that 0≤ |๐ ๐ (๐ฅ)|=| ๐(๐+1) (๐ง) (๐+1)! |๐ฅ|๐+1 = (๐+1)! ๐→∞ ๏ฎ Follows that for a fixed x lim |๐ฅ|๐+1 ๐ฅ ๐+1 | ≤ (๐+1)! (Taylor’s Theorem) 0 and by the squeeze Theorem, it follows that for all x Rn(x)๏ 0 as n๏ ∞ the Maclaurin series for sin x converges to sin x for all x. Summary : Guidelines for finding a Taylor Series: 1. Differentiate f(x) several times and evaluate each derivative at c: f(c), f’(c), f’’(c), f’’’(c)..fn(c) ๏ Try to recognize a pattern in these #’s 2. Use the sequence developed in the 1st step to for the Taylor coefficients an = determine the IOC for the resulting power series: f(c) + f’(c)(x-c) + … + ๐ (๐) (๐) ๐! ๐(๐) (๐) ๐! and (x-c)n 3. Within the IOC determine whether of not the series converges to f(x) **Direct determination of Taylor or MacLaurin coefficients using successive differentiation can be difficult ๏ below is an example of a shortcut using known coefficients of Taylor or MacLaurin Series Ex. find the MacLaurin series for f(x) = sin x2 1) To find the coefficients you must calculate successive derivatives of f(x) = sin x2 f’(x) = 2x cos x2 f’’(x) = -4x2sin x2+ 2cosx2 ๏ to keep finding derivative is cumbersome so: Alternate Method: 1) Maclaurin series for sin x found above is sin x = x - ๐ฅ3 3! + ๐ฅ5 5! - ๐ฅ7 7! +… 2) Now because sin x2 = g(x)2๏ you can substitute x2 for x in the series to obtain: Sin x2 = g(x)2 = x2 - ๐ฅ6 3! + ๐ฅ 10 5! - ๐ฅ 14 + 7! … Most practical way to find a Taylor or MacLaurin Series: 1) Develop power series from a basic list of elementary functions 2) From the list ๏ you can determine power series for other functions by +/-/x/÷ and differentiation /integration a composition with a known power series. BINOMIAL SERIES – function of the form f(x) = (1+x)k Ex. Find the Maclaurin series for f(x) = (1+x)k and determine its Radius of Convergence: Successive differentiation = f’(x) =k(1+x)k-1, f’’(x)=k(k-1)(1+x)k-2, f’’’(x)=k(k-1)(k-2)(1+x)k-3, fn(x)=k..(k-n+1)(1+x)k-n ๏ f(0)=1, f’(0)=k, f’’(0)=k(k-1), f’’’(0)=k(k-1)(k-2), fn(0)=k(k-1)…(k-n+1) Which produces the series: 1+kx + Because ๐๐+1 ๏ ๐๐ ๐(๐−1)๐ฅ 2 2 + …+ ๐(๐−1)…(๐−๐+1)๐ฅ ๐ ๐! 1 you can apply the ratio test to conclude the ROC is R=1, so the series converges to some function in the interval (-1,1) 3 Ex. Find the power series for f(x) = √1 + ๐ฅ ๏ (1+x)1/3 ๐(๐−1)๐ฅ 2 ๐(๐−1)…(๐−๐+1)๐ฅ ๐ + …+ 2 ๐! 2·5๐ฅ 3 2·5·8๐ฅ 4 - 34 4! + … which converges 33 3! 1) Use the binomial series: (1+x)k = 1+kx + 2๐ฅ 2 2) Let k = 1/3 and write: 1+1/3x + 32 2 + for -1≤ x ≤ 1 ***Know deriving Power Series from Elementary Function chart in book *** Ex.. Deriving a power series from a Basic List: Find the power series for f(x) = cos√๐ฅ 1) Use power series for cos x = 1 2 cos√๐ฅ = 1 - √๐ฅ 2! 4 + √๐ฅ 4! ๐ฅ2 2! + ๐ฅ4 4! - ๐ฅ6 6! + … and replace x by √๐ฅ to get: ๐ฅ 2! + ๐ฅ2 4! - ๐ฅ3 6! … 6 - √๐ฅ 6! = 1- this series converges for all x in the domain of cos√๐ฅ ๏ that is for all x≥0 Multiplication/Division of a power Series: Ex. find the first 3 nonzero terms in each MacLaurin Series a. ๐ ๐ฅ arctanx b. tan x