π Lecture 44 β Alternating Series; Conditional Convergence
π Overview: This lecture explores alternating seriesβinfinite series with terms that alternate in signβand establishes the conditions under which they converge. It also introduces the concepts of absolute and conditional convergence, along with the ratio test for absolute convergence, and concludes by examining power series in x and in (x-a), defining their intervals and radii of convergence.
ποΈ Topics Covered
The lecture begins with an introduction to alternating series and the Alternating Series Test (Theorem 11.7.1), followed by a theorem on error estimation when approximating sums. It then defines absolute and conditional convergence (Definition 11.7.3) and presents Theorem 11.7.4 on the convergence of absolutely convergent series and Theorem 11.7.5 (Ratio Test for Absolute Convergence). Finally, it covers power series in x and in (x-a), including Theorem 11.8.1 on their intervals of convergence and the concepts of radius and interval of convergence.
π Lecture Summary
Alternating Series; Alternating Series Test
An alternating series is one whose terms alternate in sign, typically expressed as β{k=1}^{β} (-1)^{k+1} a_k = a_1 - a_2 + a_3 - a_4 + ... where all a_k are positive. The Alternating Series Test (Theorem 11.7.1) states that such a series converges if two conditions are met: (a) the positive terms a_k form a decreasing sequence (a_1 > a_2 > a_3 > ... > a_k > ...), and (b) lim{kββ} a_k = 0.
Example: The alternating harmonic series β{k=1}^{β} (-1)^{k+1} (1/k) converges because a_k = 1/k > 1/(k+1) = a{k+1} and lim_{kββ} 1/k = 0. Notably, the ordinary harmonic series diverges, but its alternating version converges.
π‘ Why this matters: The alternating series test provides a simple way to confirm convergence for many series that have both positive and negative terms.
Error Estimation in Alternating Series
Theorem 11.7.2 provides a bound on the error when approximating the sum S of a convergent alternating series by its nth partial sum S_n. The absolute error |S - S_n| is less than the first omitted term a_{n+1}, and the sign of the error is the same as the sign of that omitted term.
Example: For the alternating harmonic series, S_7 = 1 - 1/2 + 1/3 - 1/4 + 1/5 - 1/6 + 1/7 = 319/420, and S_8 = 533/840. The true sum S (ln 2) lies between these: 533/840 < S < 319/420. The error |ln2 - S_7| < a_8 = 1/8, and |ln2 - S_8| < a_9 = 1/9.
Absolute and Conditional Convergence
Definition 11.7.3: A series β u_k converges absolutely if the series of absolute values β |u_k| converges. If β u_k converges but β |u_k| diverges, the series is conditionally convergent.
Example: The series 1 - 1/2 - 1/2^2 + 1/2^3 + 1/2^4 - 1/2^5 - ... converges absolutely because the series of absolute values is a convergent geometric series. In contrast, the alternating harmonic series converges conditionally because its absolute value series is the divergent harmonic series.
Theorem 11.7.4: If a series converges absolutely, then it converges. However, if β |u_k| diverges, no conclusion can be drawn about β u_k; it may converge conditionally (as with the alternating harmonic series) or diverge (as with -1 - 1/2 - 1/3 - ...).
Theorem 11.7.5 (Ratio Test for Absolute Convergence): For a series β u_k with nonzero terms, let Ο = lim_{kββ} |u_{k+1}/u_k|. If Ο < 1, the series converges absolutely (and hence converges). If Ο > 1 or Ο = +β, the series diverges. If Ο = 1, the test is inconclusive.
Example: The series β_{k=1}^{β} (cos k)/k^2 converges absolutely because |cos k|/k^2 β€ 1/k^2, and β 1/k^2 converges by the comparison test, so by Theorem 11.7.4, the original series converges.
Power Series in x
A power series in x is an infinite series of the form β_{k=0}^{β} c_k x^k = c_0 + c_1 x + c_2 x^2 + ... + c_k x^k + ... Theorem 11.8.1 states that for any such series, exactly one of three cases holds: (a) it converges only for x = 0; (b) it converges absolutely for all real x; or (c) it converges absolutely for all x in a finite open interval (-R, R) and diverges for |x| > R. At the endpoints x = Β±R, the series may converge absolutely, converge conditionally, or diverge.
The interval of convergence is the set of x for which the series converges, and R is the radius of convergence.
Example: For the series β_{k=1}^{β} x^k / k, applying the ratio test gives Ο = |x|. The series converges absolutely for |x| < 1 (so R = 1) and diverges for |x| > 1. At x = 1, the series becomes β 1/k (harmonic), which diverges. At x = -1, it becomes β (-1)^k / k (alternating harmonic), which converges conditionally. Thus the interval of convergence is [-1, 1).
Power Series in x-a
A power series in (x-a) is of the form β_{k=0}^{β} c_k (x - a)^k = c_0 + c_1 (x - a) + c_2 (x - a)^2 + ... Theorem 11.8.1 (generalized) states that for any such series, exactly one of three cases holds: (a) it converges only for x = a; (b) it converges absolutely for all real x; or (c) it converges absolutely for all x in a finite open interval (a - R, a + R) and diverges for |x - a| > R. At the endpoints, behavior may vary. The interval of convergence is centered at x = a.
Example: The series β_{k=0}^{β} (x - 1)^k / (k+1) has a = 1. Its interval of convergence is centered at 1.
β Key Takeaways
The Alternating Series Test provides a crucial method for determining convergence of series with alternating signs, relying on decreasing positive terms that approach zero. Error bounds for alternating series approximations are straightforward: the error is less than the first omitted term. Absolute convergence is a stronger condition than ordinary convergence, and the ratio test for absolute convergence is a powerful tool for determining convergence of series with arbitrary signs, including power series. Power series in x have an interval of convergence centered at 0 (or at a for power series in x-a), with a radius of convergence R that can be found using the ratio test; endpoints must be checked separately.
π§ Quick Revision Questions
- State the two conditions of the Alternating Series Test.
- If a convergent alternating series is approximated by its 10th partial sum, what is the bound on the absolute error?
- What is the difference between absolute convergence and conditional convergence?
- For a power series β c_k x^k, what does the ratio test (Ο = lim |u_{k+1}/u_k|) tell you about the radius of convergence?
- How is the interval of convergence different from the radius of convergence?
π Lecture 45 β Taylor and Maclaurin Series
π Overview: This lecture addresses the problem of approximating functions using polynomials near a specific point. It introduces the systematic construction of Taylor and Maclaurin polynomials, which match a function and its derivatives at a point, providing excellent local approximations. The lecture demonstrates this method for the exponential function ( e^x ) and discusses the concept of convergence to the Taylor series.
ποΈ Topics Covered
The lecture begins by stating the problem of approximating a function by a polynomial near a given point ( a ). It then derives the coefficients of a Maclaurin polynomial by matching function values and derivatives at ( a=0 ), leading to a general formula. This is illustrated with an example for ( e^x ). The discussion then extends to Taylor polynomials centered at a general point ( x=a ) and introduces the Taylor and Maclaurin series as the limit of these polynomials as the degree approaches infinity.
π Lecture Summary
One of the early applications of calculus...
The lecture opens by stating that one of the early applications of calculus was computing approximate values for functions like ( \sin x, \ln x, ) and ( e^x ) using polynomial approximations.
Problem
The central problem is: Given a function f and a point a on the x-axis, find a polynomial of specified degree that best approximates the function f in the βvicinityβ of the point a. The strategy is to choose the polynomial's coefficients so that its value and the values of its first ( n ) derivatives match those of ( f ) at the point ( a ). For simplicity, this is first done for the case ( a=0 ).
Suppose that we are interested in approximating a function f...
For a polynomial ( P(x) = c_0 + c_1x + c_2x^2 + \dots + c_n x^n ), we impose ( n+1 ) conditions at ( x=0 ): ( f(0) = P(0), f'(0) = P'(0), f''(0) = P''(0), \dots, f^{(n)}(0) = P^{(n)}(0) ). By computing the derivatives of ( P(x) ), we find: ( P(0) = c_0, P'(0) = c_1, P''(0) = 2!c_2, P'''(0) = 3!c_3, \dots, P^{(n)}(0) = n!c_n ). Thus, the coefficients are determined by: ( c_0 = f(0), c_1 = f'(0), c_2 = \frac{f''(0)}{2!}, c_3 = \frac{f'''(0)}{3!}, \dots, c_n = \frac{f^{(n)}(0)}{n!} ). This defines the Maclaurin polynomial of degree ( n ) for ( f ).
Example: Find the Maclaurin polynomials Pβ, Pβ, Pβ, Pβ, and Pβ for eΛ£.
Let ( f(x) = e^x ). All its derivatives are also ( e^x ), so ( f^{(k)}(0) = e^0 = 1 ) for all ( k ). The Maclaurin polynomials are:
- ( P_0(x) = f(0) = 1 )
- ( P_1(x) = f(0) + f'(0)x = 1 + x )
- ( P_2(x) = f(0) + f'(0)x + \frac{f''(0)}{2!}x^2 = 1 + x + \frac{x^2}{2!} )
- ( P_n(x) = 1 + x + \frac{x^2}{2!} + \dots + \frac{x^n}{n!} )
π Definition β Maclaurin polynomial of degree ( n ) for a function ( f ): ( P_n(x) = f(0) + f'(0)x + \frac{f''(0)}{2!}x^2 + \frac{f'''(0)}{3!}x^3 + \dots + \frac{f^{(n)}(0)}{n!}x^n ) β This polynomial's value and first ( n ) derivatives match ( f ) at ( x=0 ).
π Formula: ( P_n(x) = \sum_{k=0}^n \frac{f^{(k)}(0)}{k!}x^k ) β A compact way to write the Maclaurin polynomial.
π Example: For ( f(x) = e^x ), ( P_2(x) = 1 + x + \frac{x^2}{2} ). Near ( x=0 ), ( e^{0.1} \approx P_2(0.1) = 1 + 0.1 + 0.005 = 1.105 ). The actual value is 1.10517, showing a good approximation.
π‘ Why this matters: These polynomials provide a way to calculate function values to any desired accuracy near a point, which was historically crucial for creating mathematical tables.
Graphs of eΛ£ and first four Maclaurin polynomials are shown here.
The graphs of ( P_1(x), P_2(x), P_3(x) ) are virtually indistinguishable from that of ( e^x ) near the origin, so these polynomials are good approximations near the origin. However, away from the origin, they do not give a good approximation. This highlights that Maclaurin polynomials are locally accurate.
To obtain polynomial approximations of f(x) that have their best accuracy near a general point x=a...
For a general point ( x = a ), we express the polynomial in powers of ( (x-a) ): ( P(x) = c_0 + c_1(x-a) + c_2(x-a)^2 + \dots + c_n(x-a)^n ). By matching derivatives at ( x=a ), we find the coefficients: ( c_k = \frac{f^{(k)}(a)}{k!} ).
π Definition β Taylor polynomial of degree ( n ) for ( f ) about ( x=a ): ( P_n(x) = f(a) + f'(a)(x-a) + \frac{f''(a)}{2!}(x-a)^2 + \dots + \frac{f^{(n)}(a)}{n!}(x-a)^n ) β This polynomial approximates ( f ) near ( x=a ).
Taylor and Maclaurin Series
For a fixed ( x ) near ( a ), the approximation by the Taylor polynomial ( P_n(x) ) should improve as ( n ) increases, since higher-order derivatives match. It is plausible that the value of ( P_n(x) ) might converge to ( f(x) ) as ( n \to \infty ). This leads to the Taylor series for ( f ) about ( x=a ): ( f(x) = f(a) + f'(a)(x-a) + \frac{f''(a)}{2!}(x-a)^2 + \dots = \sum_{k=0}^{\infty} \frac{f^{(k)}(a)}{k!}(x-a)^k ).
π Definition β Taylor series: An infinite series representation of a function, ( f(x) = \sum_{n=0}^{\infty} \frac{f^{(n)}(a)}{n!}(x-a)^n ), where ( a ) is the center of expansion.
π Definition β Maclaurin series: The special case of a Taylor series where the center ( a = 0 ), so ( f(x) = \sum_{n=0}^{\infty} \frac{f^{(n)}(0)}{n!}x^n ).
π Example: The Maclaurin series for ( e^x ) is ( e^x = \sum_{n=0}^{\infty} \frac{x^n}{n!} ). This series converges to ( e^x ) for all real ( x ).
β Key Takeaways
The most critical concept is that Taylor and Maclaurin polynomials and series provide a systematic method for approximating functions using polynomials by matching derivatives at a single point, ensuring high accuracy near that point. The coefficients are uniquely determined by the function's derivatives at the expansion point. While Maclaurin polynomials are centered at zero, Taylor polynomials are centered at any general point ( a ). The Maclaurin polynomials for ( e^x ) are a classic example, where ( P_n(x) = 1 + x + x^2/2! + \dots + x^n/n! ). Ultimately, the limit of these polynomials as the degree goes to infinity is the Taylor or Maclaurin series, which may converge to the original function.
π§ Quick Revision Questions
- What are the conditions used to determine the coefficients of a Maclaurin polynomial of degree ( n ) for a function ( f )?
- Write the formula for the Maclaurin polynomial of degree 3 for a general function ( f ).
- What is the Maclaurin polynomial of degree 2 for ( f(x) = e^x ), and how does it approximate ( e^{0.2} )?
- How does the formula for a Taylor polynomial centered at ( x = a ) differ from a Maclaurin polynomial?
- What is the definition of the Taylor series for a function ( f ) about a point ( a )?