Finding the square root of 16.2 by hand sounds hard, but we know the square root of 16 is 4. Since 16.2 is so close to 16,16.2 should be just a little more than 4. To find how much more, notice that close to (16,4), the curve y=x is almost straight and stays very near its tangent line there. So the height of that tangent line at 16.2 makes a good estimate.
Estimate 16.2 using the tangent line to y=x at x=16
Text solution
The slope at 16Write x as x21 and use the power rule.
f′(x)=2x1
f′(16)=2⋅41=81
The tangent lineWrite the line through (16,4) with slope 81 in point-slope form, then add 4 to both sides.
L(x)=4+81(x−16)
The slope times the stepThe step from 16 to 16.2 is 0.2, so the line rises 81(0.2)=0.025. Squaring 4.025 gives 16.200625, very close to 16.2.
16.2≈L(16.2)=4+0.025=4.025
This tangent line only gives a good estimate when x stays close to 16. At x=25 it gives 5.125, but 25 is exactly 5. The line's slope stays 81, while the curve's slope keeps decreasing, down to 101 at 25, so the line gets farther and farther above the curve.
4
9
16
25
1
2
3
4
5
x
y
slope 81
y=x
5.125
5
L(25)=4+81(9)=5.125
25=5
5.125−5=0.125
The same idea works for any function that has a derivative at a point a. We start at the known value f(a) and add the slope f′(a) times the step x−a. The height on this tangent line, L(x), is called the linearization of f at a, and it needs only the value and the slope at a.
16.2≈4+81(0.2)
The linearization at a
L(x)=f(a)+f′(a)(x−a)
2
Choosing the function and the point
For 16.2, we were given the function f(x)=x and the point a=16. But often we only have a number to estimate, like 326, and we choose both ourselves. Since we're taking a cube root, we let f(x)=3x, and for a we pick the nearest input where we know f(a) and f′(a) exactly. After that, we'll estimate sin31∘, and then a value of a function we have no formula for.
Estimate 326,sin31∘, and f(3.9) using tangent lines
L(x)=f(a)+f′(a)(x−a)
1
2
3
4
x
y
26
326
y=3x
Text solution
326: a step to the leftThe nearest perfect cube is 27, so a=27,f(27)=3 and f′(27)=31⋅91=271. Since 26 is to the left of 27, the step is −1.
326≈3+271(−1)=2780≈2.963
sin31∘: the step in radiansWith a=30∘, a step of 1 gives about 1.37, more than any sine can be. The derivative of sinx is cosx only when x is in radians, so the step is 1∘=180π.
sin31∘≈21+23⋅180π≈0.5151
f(3.9): no formula neededThe tangent line needs only f(4)=2 and f′(4)=−3. The step 3.9−4=−0.1 is negative, and the negative slope times the negative step gives a positive rise.
f(3.9)≈2+(−3)(−0.1)=2.3
3
Overestimates and underestimates
The tangent line at 16 gave 16.2≈4.025. Is 4.025 bigger than 16.2, or smaller? We could square 4.025 to find out, but we can also tell from the slope of the curve, without squaring anything.
Decide whether a linear approximation is an overestimate or an underestimate from the second derivative
9
16
25
3
4
5
x
y
4.025?16.2
slope 81
y=x
Text solution
Compare the slopesThe line's slope is 81 everywhere, but the curve's slope decreases as x increases. To the right of 16, the curve climbs slower than the line. To the left, it climbs faster, so to meet the line at (16,4) it had to start below it.
f′(9)=61
f′(16)=81
f′(25)=101
The sign of f′′A decreasing slope means f′′ is negative. The curve is below its tangent line on both sides of 16, so 4.025 is an overestimate.
f′′(x)=−4x231<0
ex at 0Here f′′(x)=ex is positive, so the curve is above its tangent line L(x)=1+x on both sides, and 1.1 is an underestimate.
e0.1≈1.105>1.1
f(3.9) when f′′<0The curve is still below its tangent line, even though f is decreasing and we stepped to the left.
f(3.9)<L(3.9)=2.3
Here's y=x with its tangent line at 16. On both sides of 16, the line is above the curve.
4
9
16
25
1
2
3
4
5
x
y
slope 81
y=x
3.125
3
5.125
5
L(9)=4−87=3.125
3.125>3=9
L(25)=5.125
5.125>5=25
When f′′<0 on an interval, the slope is decreasing, and we say the curve is concave down. A tangent line to a concave-down curve gives an overestimate whether we step to the right or to the left. When f′′>0, the slope is increasing, and we say the curve is concave up. That's y=ex, which stays above its tangent line at 0.
−2
−1
1
2
3
4
x
y
1
e
2
e−1
slope 1
y=ex
f(x)=exa=0
L(x)=1+x
e0.1≈L(0.1)=1.1
f′′(x)=ex>0
e0.1≈1.105>1.1
Does moving left of the tangent point change an overestimate into an underestimate?
4
Differentials and the change in y
If x changes by a small amount, how much does y=f(x) change?
f(x+dx)−f(x)≈?
To answer that, let's give the changes names. We call the change in x,dx. The change in height along the curve y=f(x) is Δy, and the rise along the tangent line is dy. We call the line's rise dy because dxdy is the slope of the tangent line, f′(x). Slope is rise over run, so with run dx, the rise is f′(x)dx.
Rise along the tangent line
dy=f′(x)dx
Rise along the curve
Δy=f(x+dx)−f(x)
When we estimated 16.2, the step 0.2 was dx, and the rise along the tangent line, 0.025, was dy. So the linear approximation is just f(a)+dy.
x at 16:dx=0.2dy=81(0.2)=0.0254.025=4+dy
But dy is the rise along the line, not the curve. On y=x2 at x=3, with dx=1, the line rises 6 and the curve rises 7.
2
3
4
x
dx=1
dy=6
Δy=7
y=x2
y=x2x=3dx=1
dy=2(3)(1)=6
Δy=42−32=7
Δy−dy=1
Why did dy miss that 1, and what happens to the difference when dx is small?
Compare dy and Δy for y=x2 at x=3
2
3
4
x
dx=1
dy
Δy
y=x2
Text solution
Two strips and a cornerThink of x2 as the area of a square with side x. Growing the side from 3 to 3+dx adds two strips, which make dy, and a corner, which dy leaves out.
(3+dx)2−32=2⋅3dx+(dx)2
Make dx smallWith dx=0.1, the strips shrink by a factor of 10, but the corner shrinks by a factor of 100. So when dx is small, dy is almost all of Δy.
0.6+0.01=0.61
Here's the square for dx=1. The corner is the 1 that dy missed.
3
dx=1
3dx
3dx
(dx)2
3
dx=0.1
3dx
3dx
(dx)2
dx=1:6+1=7
To find dy for any function, take the derivative and multiply by dx, evaluating the derivative at the starting value x, not x+dx. For y=x2sinx, we find the derivative with the product rule.
ydy=x2sinx=(2xsinx+x2cosx)dx
5
Estimating error and relative error
Say we measure the radius of a sphere to be 10 cm, but our measurement could be off by up to 0.1 cm. How much could that error change the computed volume?
We'll find that error in cubic centimetres, and also the relative error, which is the error divided by the whole volume.
Estimate the error and relative error in a sphere's volume from an error in its radius
r=10cm
Text solution
The error is a differentialFor a square of side x, an error dx gave dA=2xdx. Here the error dr=0.1 gives dV=4πr2dr, evaluated at the measured radius r=10.
dV=4π(10)2(0.1)=40π≈126 cm3
The relative errorDividing dV by V=34πr3 leaves 3⋅rdr, and 100.1=1%. The 3 is the exponent the power rule brought down from r3.
VdV=3⋅rdr=3(1%)=3%
If we measure the circumferenceWrite the area in terms of C, then multiply its derivative by the error dC.
r=2πC→A=πr2=4πC2
The factor 4πr2 is the surface area of the sphere. Just as growing the square added two thin strips, growing the radius from r to r+dr adds a thin shell of thickness dr over the whole sphere. The volume of that shell is about its surface area times its thickness, 4πr2⋅dr.
dr
4πr2
Summary
The linearization at a point
The tangent line
L(x)=f(a)+f′(a)(x−a)
To estimate f(x) near a known point a, find the height at x on the tangent line at a. Start at the known height f(a), and add the slope f′(a) times the step x−a. The formula needs only the value and the derivative at a, so two given numbers f(a) and f′(a) are enough to estimate f(x). When f is given only through an equation, use implicit differentiation or an equation for a rate like dtdW to find the derivative, then evaluate at a to get f′(a).
Choosing the point and the step
A step to the left
326:a=27x−a=−1
An angle step in radians
sin31∘:a=30∘x−a=1∘=180π
The number you want to estimate tells you the function f. Choose a as the nearest input where you know f(a) and f′(a) exactly, like a=27 for 326. If x<a, keep the step x−a negative. For trig functions, the derivative formulas only hold in radians, so always convert an angle step from degrees to radians using 1∘=180π.
Overestimates and underestimates
y=x
y=ex
When f′′(x)<0 near a, the curve is concave down and lies below its tangent line on both sides of a, so L(x) is an overestimate. When f′′(x)>0 near a, the curve is concave up and lies above its tangent line on both sides of a, so L(x) is an underestimate. Whether the step goes to the left or to the right doesn't change the result.
Differentials and the actual change
3
dx
3dx
3dx
(dx)2
Along the tangent line and along the curve
dyΔy=f′(x)dx=f(x+dx)−f(x)
dx is the change in x. Over that run, dy=f′(x)dx is the rise along the tangent line, while Δy=f(x+dx)−f(x) is the actual rise along the curve. For a square of side 3 growing by dx, the two side strips have total area 2⋅3dx, which is dy. The corner has area (dx)2, which is the difference Δy−dy. When dx is small, the corner shrinks much faster than the strips, so dy≈Δy.
Estimating measurement error
Error in volume
dV=4πr2drVdV=3⋅rdr
Write the calculated quantity in terms of the measured value x, and treat the measurement error as dx. The estimated error in the quantity is the differential df=f′(x)dx, evaluated at the measured value. The relative error is the estimated error divided by the calculated value, f(x)df. For a sphere, the power rule brings down the exponent 3, so a 1% error in the radius gives about a 3% error in the volume.
6
Practice
Give decimals for estimates, and exact values where they exist.
1
Find the linearization L(x) of f(x)=x32 at a=8, and use it to estimate (8.06)32.
To estimate (8.06)32, what nearby number has an exact cube root?
The power rule gives f′(8)=32⋅8−31. What is 8−31?
If you got 4.08 Check 8−31. The negative exponent means 381, which is 21, not 2.
The value at 8The cube root of 8 is 2, and squaring gives 4.
f(8)=832=(38)2=4
The derivative at 8The power rule gives 32x−31, and 8−31 is 381=21.
f′(x)=32x−31,f′(8)=32⋅21=31
Put both into the linearization L(x)=f(a)+f′(a)(x−a).
L(x)=4+31(x−8)
Plug in x=8.06, so the step x−8 is 0.06.
(8.06)32≈L(8.06)=4+31(0.06)=4.02
2
(a) Use a linear approximation to estimate (1.02)10. (b) Is your estimate an overestimate or an underestimate? Explain.
For f(x)=x10, what starting point a is close to 1.02?
If you got 1.02 Don't forget to multiply the step x−a=0.02 by the derivative f′(1)=10.
(b)
Find f′′(x). Is it positive or negative near 1?
Linearization at 1We choose a=1 because powers of 1 are easy, and 1.02 is only 0.02 away.
f(x)=x10,a=1f(1)=1,f′(x)=10x9,f′(1)=10
Multiply the step 0.02 by the derivative 10, then add the starting value 1.
(1.02)10≈1+10(0.02)=1.2
Check the concavityThe second derivative 90x8 is positive near 1, so the curve is concave up and the tangent line lies below it.
f′′(x)=90x8>0→(1.02)10>1.2
The true value is about 1.219, so 1.2 is an underestimate.
(1.02)10≈1.219
3
The curve x2+xy+y2=7 passes through the point (1,2). Use the tangent line at (1,2) to estimate y when x=1.1.
Differentiate both sides with respect to x to find dxdy at (1,2).
The starting y is 2 and the step in x is 0.1. How does y change along the tangent line?
If you got 2.08 Check the sign of dxdy. Solving 4+5dxdy=0 gives a negative slope.
Differentiate implicitlyDifferentiate both sides with respect to x, using the product rule on xy and the chain rule on y2.
2x+y+xdxdy+2ydxdy=0→dxdy=−x+2y2x+y
Plug in x=1 and y=2 to find the slope of the tangent line.
(1,2):dxdy=−1+42+2=−54
The step in x is 1.1−1=0.1, so multiply the slope by 0.1 and add the starting y-value.
y(1.1)≈2−54(0.1)=2−0.08=1.92
4
Let y=x3−2x. Find dy. Then compute dy and Δy when x=2 and dx=0.1.
The differential dy is f′(x) times what?
Δy is the actual change f(x+dx)−f(x). What are x+dx and x?
If you got dy=1010 is the derivative f′(2). Multiply by dx=0.1 to find dy.
The differential dy is the derivative times dx.
dy=(3x2−2)dx
Plug in x=2 to get the slope 10, then multiply by dx=0.1.
x=2,dx=0.1:dy=(3⋅4−2)(0.1)=10(0.1)=1
The actual change Δy is the change on the curve from x=2 to x=2.1.
Δy=f(2.1)−f(2)=(9.261−4.2)−(8−4)=5.061−4=1.061
5
The circumference of a circle is measured as 84 cm, with a possible error of 0.5 cm. (a) Use differentials to estimate the error in the computed area. (b) Estimate the relative error in the area.
The circumference C was measured, so write the area A in terms of C.
If you got 42 The measurement error 0.5 cm is dC, not dr. Write A in terms of C before differentiating.
The relative error in area is AdA. How does it simplify in terms of CdC?
Area in terms of circumferenceWrite the area directly in terms of C because the measurement error 0.5 cm is dC.
r=2πC→A=πr2=4πC2
Error in areaDifferentiate with respect to C, then plug in C=84 and dC=0.5.
dA=2πCdC=2π84(0.5)=π21≈6.7 cm2
Relative errorDivide dA by A. Because A is proportional to C2, the relative error in A is twice the relative error in C.
AdA=4πC22πCdC=2⋅CdC=2⋅840.5=841≈1.2%
6
A hemispherical dome 50 m across gets a coat of paint 0.05 cm thick. Use differentials to estimate the amount of paint needed.
The paint is a thin shell over the dome. What is the radius of the dome if it is 50 m across?
The thickness is dr=0.05 cm. How many meters is that?
If you got 62.5π Check your units. The thickness is 0.05 cm, which is 0.0005 m, not 0.05 m.If you got 2.5π Check the radius. The dome is 50 m across, so the radius is 25 m, not 50 m.
The differential of volumeThe paint forms a thin shell over the hemisphere, so its volume is approximately the surface area 2πr2 times the thickness dr.
V=32πr3→dV=2πr2dr
Radius and thicknessThe radius is half the diameter, and dividing 0.05 cm by 100 converts the thickness to meters.
r=250=25 m,dr=0.05 cm=0.0005 m
Plug r=25 and dr=0.0005 into dV.
dV=2π(25)2(0.0005)=0.625π≈1.96 m3
7
(Challenge) At the start of 2010 a landfill holds 1400 tons of waste. The amount W, in tons, t years after the start of 2010, is increasing and satisfies dtdW=25W−300. (a) Use the tangent line to the graph of W at t=0 to estimate the amount of waste at the end of the first 3 months of 2010. (b) Find dt2d2W in terms of W, and use it to decide whether your estimate in (a) is an underestimate or an overestimate.
Find W′(0) from the differential equation. What fraction of a year is 3 months?
If you got 1532 Time t is measured in years, so 3 months is 41 of a year, not 3 years.
(b)
Differentiate dtdW=25W−300 with respect to t using the chain rule.
Find the initial rate of changePlug W=1400 into the differential equation to find W′(0).
W′(0)=251400−300=44
Estimate the waste at 3 monthsThree months is 41 of a year, so multiply the slope 44 by 41 and add the starting amount 1400.
3 months=41 year:W(41)≈1400+44⋅41=1411 tons
Find the second derivativeDifferentiate 25W−300 with respect to t using the chain rule, and substitute dtdW back in.
dt2d2W=251dtdW=625W−300
Since W starts at 1400 and is increasing, W−300 is always positive, so the graph is concave up and the tangent line gives an underestimate.