10

Day 10

Related Rates


Exponential Growth & Decay

Whenever a quantity's rate of change is proportional to the quantity itself — population, radioactive mass, temperature difference, money under continuous interest — it follows the exact same exponential model. Recognizing this pattern turns a wide variety of word problems into the same three-step solve.

Core Theorem
If dydt=ky\frac{dy}{dt} = ky for a constant kk, then y=y0ekty = y_0 e^{kt}, where y0=y(0)y_0 = y(0).
k>0k>0 gives exponential growth; k<0k<0 gives exponential decay.
Half-life
TT satisfies 12=ekT\frac{1}{2} = e^{kT}, i.e. k=ln2Tk = -\frac{\ln 2}{T}.
Newton's Law of Cooling:
H(t)=Hs+(H0Hs)ektH(t) = H_s + (H_0-H_s)e^{-kt}, where HsH_s is the surrounding temperature and H0H_0 is the initial temperature.
Watch the TikTok ExplanationExponential Growth and Decay
Step-by-Step SOP
  1. 1

    Identify y₀ and Set Up y = y₀e^(kt)

    Find the initial value (at t=0t=0) to use as y0y_0; if the problem gives a half-life or doubling time instead, convert that to kk directly.
  2. 2

    Solve for k Using the Given Data Point

    Substitute the known (t,y)(t,y) pair and isolate kk using logarithms.
  3. 3

    Answer the Actual Question

    Substitute back into the full formula to evaluate at a new time, or solve for tt given a target value of yy.

Practice Exercises


Example 01Medium
World population was 2.562.56 billion in 1950 and 3.043.04 billion in 1960. Assuming exponential growth, find the relative growth rate, then estimate the population in 1993 (43 years after 1950).
NEED A HINT?
Let t=0t=0 correspond to 1950. Use the 1960 data point (t=10t=10) to solve for kk first.
SHOW DETAILED EXPLANATION

Step 1: Set Up the Model

P(t)=2.56ektP(t) = 2.56e^{kt} (billions), with tt in years since 1950.

Step 2: Solve for k Using the 1960 Data

P(10)=3.04=2.56e10k    e10k=3.042.56=1.1875    k=ln(1.1875)100.0172 per yearP(10) = 3.04 = 2.56e^{10k} \implies e^{10k} = \frac{3.04}{2.56} = 1.1875 \implies k = \frac{\ln(1.1875)}{10} \approx 0.0172\ \text{per year}.

Step 3: Estimate the 1993 Population (t = 43)

P(43)=2.56e43(0.0172)=2.56e0.7392.56(2.094)5.36P(43) = 2.56e^{43(0.0172)} = 2.56e^{0.739} \approx 2.56(2.094) \approx 5.36 billion — reasonably close to the true 1993 figure of about 5.55.5 billion.
Example 02Medium
Radium-226 has a half-life of 15901590 years. A sample starts at 100mg100\text{mg}. Find the mass after 10001000 years, and the time until only 30mg30\text{mg} remains.
NEED A HINT?
Use the half-life to find kk first: 12=e1590k\frac{1}{2} = e^{1590k}.
SHOW DETAILED EXPLANATION

Step 1: Find k from the Half-Life

k=ln215900.000436 per yeark = -\frac{\ln 2}{1590} \approx -0.000436\ \text{per year}.

Step 2: Write the Mass Function

m(t)=100e0.000436tm(t) = 100e^{-0.000436t}.

Step 3: Evaluate at t = 1000

m(1000)=100e0.436100(0.6466)64.7mgm(1000) = 100e^{-0.436} \approx 100(0.6466) \approx 64.7\text{mg}.

Step 4: Solve for t When m(t) = 30

30=100e0.000436t    e0.000436t=0.3    t=ln(0.3)0.0004362762 years30 = 100e^{-0.000436t} \implies e^{-0.000436t} = 0.3 \implies t = \frac{\ln(0.3)}{-0.000436} \approx 2762\ \text{years}.
Example 03Hard
A cup of water at 22°C22°C is placed in a 7°C7°C refrigerator. After half an hour, its temperature is 16°C16°C. Find the temperature after one full hour, and the time needed to reach 10°C10°C.
NEED A HINT?
Use H(t)=Hs+(H0Hs)ektH(t) = H_s + (H_0-H_s)e^{-kt} with Hs=7H_s=7, H0=22H_0=22, and the half-hour data point to solve for kk.
SHOW DETAILED EXPLANATION

Step 1: Set Up the Cooling Equation

H(t)=7+15ektH(t) = 7 + 15e^{-kt}.

Step 2: Solve for k Using H(0.5) = 16

16=7+15e0.5k    e0.5k=915=0.6    k=2ln(0.6)1.022 per hour16 = 7+15e^{-0.5k} \implies e^{-0.5k} = \frac{9}{15} = 0.6 \implies k = -2\ln(0.6) \approx 1.022\ \text{per hour}.

Step 3: Find H(1)

Since ek=(e0.5k)2=0.62=0.36e^{-k} = (e^{-0.5k})^2 = 0.6^2 = 0.36: H(1)=7+15(0.36)=12.4°CH(1) = 7+15(0.36) = 12.4°C.

Step 4: Solve for t When H(t) = 10

10=7+15e1.022t    e1.022t=0.2    t=ln(0.2)1.0221.575 hours94.5 minutes10 = 7+15e^{-1.022t} \implies e^{-1.022t} = 0.2 \implies t = \frac{\ln(0.2)}{-1.022} \approx 1.575\ \text{hours} \approx 94.5\ \text{minutes}.
Example 04Easy
You invest 10001000 dollars at a continuously compounded interest rate of 6%6\%. Find the amount after 3 years.
NEED A HINT?
Continuous compounding is exactly the exponential growth model: A(t)=A0ertA(t) = A_0e^{rt}.
SHOW DETAILED EXPLANATION

Step 1: Set Up the Formula

A(t)=1000e0.06tA(t) = 1000e^{0.06t}.

Step 2: Evaluate at t = 3

A(3)=1000e0.181000(1.1972)1197.22A(3) = 1000e^{0.18} \approx 1000(1.1972) \approx 1197.22 dollars.
Example 05Medium
A quantity yy satisfies y(t)=ky(t)y'(t) = ky(t) with y(0)=5y(0)=5 and y(1)=15y(1)=15. Find kk and y(2)y(2).
NEED A HINT?
Write y(t)=5ekty(t)=5e^{kt}, use y(1)=15y(1)=15 to solve for kk exactly (the answer will be a clean logarithm).
SHOW DETAILED EXPLANATION

Step 1: Set Up and Solve for k

y(t)=5ekty(t) = 5e^{kt}. y(1)=15    5ek=15    ek=3    k=ln3y(1)=15 \implies 5e^k=15 \implies e^k=3 \implies k=\ln 3.

Step 2: Find y(2)

y(2)=5e2ln3=5(eln3)2=5(32)=5(9)=45y(2) = 5e^{2\ln 3} = 5(e^{\ln 3})^2 = 5(3^2) = 5(9) = 45.
Common Pitfalls
  • Always Solve for k FirstEvery one of these problems follows the same order: use the given data point to pin down kk exactly, THEN answer whatever the question actually asks — trying to skip straight to the answer usually backfires.
  • Keep Units ConsistentIf tt is measured in years for one part of a problem, don't accidentally switch to months or hours later without converting — this is a very common careless error.

Linearization & Differentials

Near any point where a function is differentiable, its curve looks almost exactly like its tangent line — close enough to use that line as a fast approximation tool, and to estimate how errors in an input propagate into errors in the output.

Core Theorem
The linearization of ff at x=ax=a is L(x)=f(a)+f(a)(xa)L(x) = f(a) + f'(a)(x-a), and f(x)L(x)f(x) \approx L(x) for xx near aa.
The differential
dy=f(x)dxdy = f'(x)\,dx approximates the actual change Δy=f(x+Δx)f(x)\Delta y = f(x+\Delta x) - f(x) when Δx\Delta x (written dxdx) is small.
Step-by-Step SOP
  1. 1

    Choose a Nearby 'Nice' Point a

    Pick a point close to the value you actually want, where f(a)f(a) and f(a)f'(a) are easy to compute exactly.
  2. 2

    Build L(x) = f(a) + f'(a)(x-a)

    Compute f(a)f(a) and f(a)f'(a), then substitute into the linearization formula.
  3. 3

    For Error Propagation, Use dy = f'(x)dx

    Treat the measurement error as dxdx, and the resulting output error is approximately f(x)dxf'(x)\,dx.

Practice Exercises


Example 01Easy
Find the linearization of f(x)=3+xf(x) = \sqrt{3+x} at x=0x=0 and at x=1x=1.
NEED A HINT?
Compute f(a)f(a) and f(a)f'(a) at each point, then plug into L(x)=f(a)+f(a)(xa)L(x)=f(a)+f'(a)(x-a).
SHOW DETAILED EXPLANATION

Step 1: Find f'(x)

f(x)=123+xf'(x) = \frac{1}{2\sqrt{3+x}}.

Step 2: Linearize at a = 0

f(0)=3f(0)=\sqrt{3}, f(0)=123f'(0)=\frac{1}{2\sqrt3}. L(x)=3+123xL(x) = \sqrt{3} + \frac{1}{2\sqrt{3}}x.

Step 3: Linearize at a = 1

f(1)=2f(1)=2, f(1)=14f'(1)=\frac{1}{4}. L(x)=2+14(x1)L(x) = 2 + \frac{1}{4}(x-1).
Example 02Medium
Use linearization to estimate 3.98\sqrt{3.98} and 4.05\sqrt{4.05}.
NEED A HINT?
Pick a=4a=4 since 4=2\sqrt{4}=2 is clean, and 3.983.98, 4.054.05 are both very close to it.
SHOW DETAILED EXPLANATION

Step 1: Set Up the Linearization at a = 4

f(x)=xf(x)=\sqrt{x}, f(x)=12xf'(x)=\frac{1}{2\sqrt{x}}. f(4)=2f(4)=2, f(4)=14f'(4)=\frac{1}{4}. L(x)=2+14(x4)L(x) = 2+\frac{1}{4}(x-4).

Step 2: Estimate √3.98

L(3.98)=2+14(0.02)=20.005=1.995L(3.98) = 2+\frac{1}{4}(-0.02) = 2-0.005 = 1.995.

Step 3: Estimate √4.05

L(4.05)=2+14(0.05)=2+0.0125=2.0125L(4.05) = 2+\frac{1}{4}(0.05) = 2+0.0125 = 2.0125.
Example 03Hard
For f(x)=x3+x22x+1f(x) = x^3+x^2-2x+1, compare Δy\Delta y and dydy as xx changes from 22 to 2.052.05, and separately from 22 to 2.012.01.
NEED A HINT?
Compute dy=f(2)Δxdy=f'(2)\Delta x for each case, then compute the actual Δy=f(2+Δx)f(2)\Delta y = f(2+\Delta x)-f(2) and compare.
SHOW DETAILED EXPLANATION

Step 1: Find f'(2)

f(x)=3x2+2x2    f(2)=12+42=14f'(x)=3x^2+2x-2 \implies f'(2) = 12+4-2=14. Also f(2)=8+44+1=9f(2)=8+4-4+1=9.

Step 2: Case Δx = 0.05

dy=14(0.05)=0.7dy = 14(0.05)=0.7. f(2.05)=2.053+2.0522(2.05)+19.7176f(2.05) = 2.05^3+2.05^2-2(2.05)+1 \approx 9.7176, so Δy0.7176\Delta y \approx 0.7176 — close to dydy, but not exact.

Step 3: Case Δx = 0.01

dy=14(0.01)=0.14dy=14(0.01)=0.14. f(2.01)9.1407f(2.01) \approx 9.1407, so Δy0.1407\Delta y \approx 0.1407 — much closer to dydy than in the previous case.

Step 4: Conclude

As Δx\Delta x shrinks, dydy becomes a better and better approximation of the true change Δy\Delta y — this is exactly the content of the linear approximation theorem.
Example 04Medium
A sphere's radius is measured as 21cm21\text{cm}, with a possible measurement error of 0.05cm0.05\text{cm}. Estimate the resulting error in the computed volume.
NEED A HINT?
Use the differential dV=4πr2drdV = 4\pi r^2\,dr with r=21r=21 and dr=0.05dr=0.05.
SHOW DETAILED EXPLANATION

Step 1: Differentiate V(r)

V=43πr3    dV=4πr2drV = \frac{4}{3}\pi r^3 \implies dV = 4\pi r^2\,dr.

Step 2: Substitute

dV=4π(21)2(0.05)=4π(441)(0.05)=88.2π277.1 cm3dV = 4\pi(21)^2(0.05) = 4\pi(441)(0.05) = 88.2\pi \approx 277.1\ \text{cm}^3.

Step 3: Interpret

A tiny 0.05cm0.05\text{cm} error in the radius measurement could throw off the computed volume by roughly 277 cm3277\ \text{cm}^3.
Example 05Easy
Use linearization to estimate 8.13\sqrt[3]{8.1}.
NEED A HINT?
Use a=8a=8 since 83=2\sqrt[3]{8}=2 is clean.
SHOW DETAILED EXPLANATION

Step 1: Set Up the Linearization

f(x)=x1/3f(x)=x^{1/3}, f(x)=13x2/3f'(x)=\frac{1}{3}x^{-2/3}. f(8)=2f(8)=2, f(8)=13(8)2/3=112f'(8)=\frac{1}{3}(8)^{-2/3}=\frac{1}{12}. L(x)=2+112(x8)L(x)=2+\frac{1}{12}(x-8).

Step 2: Estimate

L(8.1)=2+112(0.1)2.0083L(8.1) = 2+\frac{1}{12}(0.1) \approx 2.0083.
Common Pitfalls
  • Linearization Only Works Near aL(x)L(x) is a good approximation only for xx close to aa — the farther away you estimate, the worse the approximation gets.
  • dx and Δx Are the Same ThingIn this context, dxdx is just notation for the (small) change in xx — it's identical to Δx\Delta x, unlike dydy and Δy\Delta y, which are only approximately equal.
  • dy Approximates Δy, But Isn't ExactThe differential dydy is a linear estimate — it will always differ slightly from the true Δy\Delta y, and that gap shrinks as Δx0\Delta x \to 0 (see Example 3).
Gary Chang

Gary Chang

Calculus Educator

5+ Years of Calculus Teaching Experience | AP Calculus Specialist. Dedicated to helping students master calculus through step-by-step logic and clear visualizations.