2.3 - Maclaurin expansions

2.3 - Maclaurin expansions

Derivatives at zero can build a polynomial that follows a function near zero. You will derive initial Maclaurin coefficients, obtain them by successive implicit differentiation, and retain enough terms when using an expansion for an approximation or a limit.

Derivatives at zero

Can information at x=0x=0 predict the value of e0.2e^{0.2}? A polynomial built from a few derivatives of e2xe^{2x} will give a close approximation at x=0.1x=0.1.

Suppose a local expansion begins

f(x)=a0+a1x+a2x2+a3x3+⋯ .f(x)=a_0+a_1x+a_2x^2+a_3x^3+\cdots.

At zero, all positive powers vanish, leaving a0=f(0)a_0=f(0). Differentiate and set x=0x=0 again: a1=f′(0)a_1=f'(0). Repeating gives 2a2=f′′(0)2a_2=f''(0) and 6a3=f′′′(0)6a_3=f'''(0). The factors 22 and 66 come from differentiating a power down to a constant.

Maclaurin series

A Maclaurin series is a power-series expansion about x=0x=0. Its initial coefficients are determined by derivatives evaluated at zero. Truncating it after a few terms gives a polynomial approximation near zero, wherever the expansion represents the function.

It is the coefficient of a power, rather than the derivative alone, that goes into the expansion.

Initial Maclaurin terms

For a function with a Maclaurin expansion,

f(x)=f(0)+xf′(0)+x22!f′′(0)+x33!f′′′(0)+x44!f(4)(0)+⋯ .\begin{aligned} f(x)={}&f(0)+xf'(0)+\frac{x^2}{2!}f''(0) +\frac{x^3}{3!}f'''(0)\\ &+\frac{x^4}{4!}f^{(4)}(0)+\cdots. \end{aligned}

Here 2!=22!=2, 3!=63!=6 and 4!=244!=24. For a truncated approximation, replace the equality and omitted terms by ≈\approx.

The Maclaurin formula is in the formula booklet. Use it to calculate the required initial coefficients.

A zero derivative gives a zero coefficient. Keep that place in your calculation so that the remaining powers do not shift.

Worked example 1: an exponential with a scaled argument

Question. Find the Maclaurin expansion of f(x)=e2xf(x)=e^{2x} up to and including the term in x3x^3.

Plan. Differentiate three times, evaluate at zero, then divide the higher derivatives by their factorials.

The chain rule gives

f′=2e2x,f′′=4e2x,f′′′=8e2x.f'=2e^{2x},\qquad f''=4e^{2x},\qquad f'''=8e^{2x}.

At zero, f(0)=1f(0)=1, f′(0)=2f'(0)=2, f′′(0)=4f''(0)=4 and f′′′(0)=8f'''(0)=8. Hence

e2x=1+2x+42x2+86x3+⋯ .e^{2x}=1+2x+\frac4{2}x^2+\frac8{6}x^3+\cdots.

Answer. e2x=1+2x+2x2+43x3+⋯\displaystyle e^{2x}=1+2x+2x^2+\frac43x^3+\cdots.

Check. At x=0.1x=0.1, the polynomial gives 1.221333333…1.221333333\ldots, while direct evaluation gives e0.2=1.221402758…e^{0.2}=1.221402758\ldots. The small difference, about 0.00006940.0000694, is consistent with an approximation rather than an exact cubic identity.

Change only the sign of the argument. Odd derivatives change sign, while even derivatives retain theirs.

Worked example 2: a negative inner scale

Question. Find the Maclaurin expansion of f(x)=e−2xf(x)=e^{-2x} up to and including the term in x3x^3.

Plan. Retain the factor −2-2 at every differentiation before evaluating at zero.

f′=−2e−2x,f′′=4e−2x,f′′′=−8e−2x.f'=-2e^{-2x},\qquad f''=4e^{-2x},\qquad f'''=-8e^{-2x}.

Substituting x=0x=0 gives the values 1,−2,4,−81,-2,4,-8. Therefore

e−2x=1−2x+42x2−86x3+⋯ .e^{-2x}=1-2x+\frac4{2}x^2-\frac8{6}x^3+\cdots.

Answer. e−2x=1−2x+2x2−43x3+⋯\displaystyle e^{-2x}=1-2x+2x^2-\frac43x^3+\cdots.

Check. Direct evaluation at x=0.1x=0.1 gives e−0.2=0.818730753…e^{-0.2}=0.818730753\ldots; the cubic gives 0.818666667…0.818666667\ldots. Also, multiplying the two example polynomials gives constant term 11 and zero coefficients of xx, x2x^2 and x3x^3, agreeing with e2xe−2x=1e^{2x}e^{-2x}=1 to the calculated degree.

The formula booklet also contains standard series, including the series for exe^x. Substituting 2x2x or −2x-2x into that supplied series independently reproduces these expansions. When a question asks for a derivation by differentiation, show the derivative values instead of merely quoting a series.

Watch out

The coefficient of x3x^3 is f′′′(0)/6f'''(0)/6, not f′′′(0)f'''(0). For e−2xe^{-2x}, forgetting the factorial would incorrectly give −8x3-8x^3 instead of −4x3/3-4x^3/3.

Use a fresh inner scale and show the derivative values.

Implicit differentiation for coefficients

A derivative relation can generate the values needed for an expansion without producing long explicit formulae for every higher derivative. Tangent is a useful example.

If y=tan⁡xy=\tan x, then

y′=sec⁡2x=1+tan⁡2x=1+y2.y'=\sec^2x=1+\tan^2x=1+y^2.

Differentiate this relation while treating yy as a function of xx:

y′′=2yy′,y′′′=2(y′)2+2yy′′.y''=2yy', \qquad y'''=2(y')^2+2yy''.

These are the same product and chain rules used for implicit second derivatives. To evaluate at zero, first find y(0)y(0), then y′(0)y'(0), then y′′(0)y''(0), and continue in that order. Each new value depends on values already found.

Problem-solving

Simplify a first derivative into a relation involving xx and yy where possible. Successive differentiation of that relation often gives the initial Maclaurin coefficients with much less algebra.

The first example uses this relation to derive, rather than assume, the tangent expansion.

Worked example 3: tangent through the cubic term

Question. Starting from y=tan⁡xy=\tan x, derive its Maclaurin expansion up to and including the term in x3x^3.

Plan. Use y′=1+y2y'=1+y^2 and its successive derivatives to evaluate the needed values at zero.

Since y(0)=0y(0)=0, the derivative relation gives

y′(0)=1+02=1y′′(0)=2y(0)y′(0)=0y′′′(0)=2(y′(0))2+2y(0)y′′(0)=2.\begin{aligned} y'(0)&=1+0^2=1\\ y''(0)&=2y(0)y'(0)=0\\ y'''(0)&=2\bigl(y'(0)\bigr)^2+2y(0)y''(0)=2. \end{aligned}

The Maclaurin formula now gives

tan⁡x=0+x+02x2+26x3+⋯ .\tan x=0+x+\frac02x^2+\frac26x^3+\cdots.

We are working near zero, on the central tangent branch −π/2<x<π/2-\pi/2<x<\pi/2.

Answer. tan⁡x=x+13x3+⋯\displaystyle \tan x=x+\frac13x^3+\cdots, with zero constant and quadratic coefficients.

Check. Multiply the proposed polynomial by the supplied initial cosine terms:

(x+x33)(1−x22)=x−x36+terms of higher degree.\left(x+\frac{x^3}3\right)\left(1-\frac{x^2}2\right) =x-\frac{x^3}6+\text{terms of higher degree}.

These are the initial sine terms, agreeing with tan⁡xcos⁡x=sin⁡x\tan x\cos x=\sin x.

Change the argument from xx to 2x2x. The first-derivative relation acquires a factor of 22, and that factor affects every later derivative value.

Worked example 4: tangent with a scaled argument

Question. For y=tan⁡(2x)y=\tan(2x), use successive implicit differentiation to find the Maclaurin expansion through the term in x3x^3.

Plan. Start with y′=2(1+y2)y'=2(1+y^2), then differentiate this relation twice.

The chain and product rules give

y′′=4yy′,y′′′=4(y′)2+4yy′′.y''=4yy', \qquad y'''=4(y')^2+4yy''.

At zero, y=0y=0. Hence

y′(0)=2,y′′(0)=0,y′′′(0)=4(2)2=16.y'(0)=2,\qquad y''(0)=0,\qquad y'''(0)=4(2)^2=16.

Substitute into the Maclaurin formula:

y=2x+166x3+⋯ .y=2x+\frac{16}{6}x^3+\cdots.

The local calculation is on the central branch −π/4<x<π/4-\pi/4<x<\pi/4.

Answer. tan⁡(2x)=2x+83x3+⋯\displaystyle \tan(2x)=2x+\frac83x^3+\cdots.

Check. Multiplying 2x+8x3/32x+8x^3/3 by 1−2x21-2x^2, the cosine expansion for cos⁡(2x)\cos(2x), gives 2x−4x3/32x-4x^3/3 through degree 33. This agrees with the sine expansion for sin⁡(2x)\sin(2x), so the quotient is consistent.

Continue this lesson.

You’re previewing NovaLearn without an account. Create a free account to read the rest.

Create free account