2.9 · AHL · HL onlyPaper 1Paper 2Paper 3

Further function models

In a minute

AHL adds , a+bln⁡xa + b\ln x, sine waves in with a , growth that levels off, and piecewise models whose pieces join up. Each time: find the , use your GDC, and say what the numbers mean.

You'll be able to

  • Find the half-life of an model and use it to predict.
  • Use natural logarithmic models f(x)=a+bln⁡xf(x) = a + b\ln x, find aa and bb from data, and interpret them.
  • Use sinusoidal models f(x)=asin⁡(b(x−c))+df(x) = a\sin(b(x - c)) + d in radians, with 2πb\frac{2\pi}{b} and phase shift cc.
  • Use logistic models, find their parameters, and interpret the .
  • Find a parameter that makes a piecewise model .
  • Use and interpret a model the question gives you.
Before you start3 quick questions on what this builds on
Check yourself
Solve ex=5\mathrm{e}^{x} = 5. Give xx to 3 s.f.
Check yourself
Which is the horizontal of y=50(0.7)t+12y = 50(0.7)^{t} + 12?
Check yourself
Write 45∘45^{\circ} in radians.

Revise first: 1.5 Exponents and introductory logarithms, 1.13 Complex numbers in polar and exponential form, 2.2 Function concepts, domain and range, 2.4 Key features of graphs, 2.5 Modelling with functions, 2.6 Fitting and testing models, 2.8 Transformations of graphs and 3.7 Radians, arc length and sector area

Half-life

In exponential decay towards 00, each equal step in time multiplies the amount by the same factor. So the time to halve is always the same, whatever the starting amount.

Learn this
N(t)=N0e−kt, k>0  ⇒  half-life=ln⁡2kN(t) = N_{0}\mathrm{e}^{-kt},\ k > 0 \;\Rightarrow\; \text{half-life} = \frac{\ln 2}{k}

Why: in N0e−kt=12N0N_{0}\mathrm{e}^{-kt} = \tfrac{1}{2}N_{0} the N0N_{0} cancels, so t=ln⁡2kt = \frac{\ln 2}{k}. It is not in the formula booklet.

If you know the half-life TT instead, the same model is N(t)=N0(12)t/TN(t) = N_{0}\left(\tfrac{1}{2}\right)^{t/T}.

Worked exampleThe mass of caffeine in a person's body, CC mg, tt hours after a coffee is modelled by C(t)=95e−0.13tC(t) = 95\mathrm{e}^{-0.13t}. Find the half-life of caffeine in the body, to 3 s.f., and interpret it.

Try it first, then reveal one step at a time.

  1. Half of the starting 9595 mg is 47.547.5 mg. Solve 95e−0.13t=47.595\mathrm{e}^{-0.13t} = 47.5.
Worked exampleA hospital receives 4040 mg of a radioactive tracer. After 99 hours, 31.531.5 mg is left. The mass, mm mg, after tt hours is modelled by m(t)=40e−ktm(t) = 40\mathrm{e}^{-kt}. (a) Use your GDC to find kk. (b) Find the half-life, to 3 s.f. (c) Find how long it takes until 55 mg is left.

Try it first, then reveal one step at a time.

  1. (a) Write down the equation: 40e−9k=31.540\mathrm{e}^{-9k} = 31.5. The solver gives k=0.026543…k = 0.026543\ldots Store the full value.
Check yourself
A medicine has a half-life of 66 hours in the blood. A patient takes 200200 mg. Find the amount left after 1515 hours, in mg, to 3 s.f.
Check yourself
A sample of 8080 g of a radioactive substance has a half-life of 55 days. What is the half-life of a 2020 g sample of the same substance?

Natural logarithmic models

A young tree grows fast, then more slowly, but it keeps growing. Doubling its age adds the same height whether it is 22 or 2020 years old. A natural logarithm model behaves like this.

ln⁡1=0\ln 1 = 0, so a=f(1)a = f(1): the value when x=1x = 1, not when x=0x = 0. The model has no value at x=0x = 0, and the yy-axis is a vertical asymptote.

ln⁡(mx)=ln⁡m+ln⁡x\ln(mx) = \ln m + \ln x, so multiplying xx by mm adds bln⁡mb\ln m to ff. With b>0b > 0 the graph keeps rising, more and more slowly, and has no horizontal asymptote.

Worked exampleThe height, hh m, of a tree tt years after it is planted is modelled by h(t)=1.2+2.5ln⁡th(t) = 1.2 + 2.5\ln t, t≥1t \ge 1. (a) Find the height after 66 years. (b) Interpret the value 1.21.2. (c) Find when the tree reaches 77 m, to 3 s.f.

Try it first, then reveal one step at a time.

  1. (a) h(6)=1.2+2.5ln⁡6=5.679…h(6) = 1.2 + 2.5\ln 6 = 5.679\ldots, about 5.685.68 m.
Worked exampleA farm trial models the yield of wheat, YY tonnes per hectare, against the fertiliser used, xx kg per hectare, by Y=a+bln⁡xY = a + b\ln x. With 2020 kg the yield is 3.13.1 tonnes; with 8080 kg it is 4.64.6 tonnes. (a) Find aa and bb. (b) Predict the yield with 150150 kg, to 3 s.f. (c) Comment on the model for small xx.

Try it first, then reveal one step at a time.

  1. (a) Substitute both points: a+bln⁡20=3.1a + b\ln 20 = 3.1 and a+bln⁡80=4.6a + b\ln 80 = 4.6.
Check yourself
A games company models a player's skill score, SS, after hh hours of practice by S=35+18ln⁡hS = 35 + 18\ln h, h≥1h \ge 1. Find the number of hours of practice for a score of 100100, to 3 s.f.
Check yourself
For the model S=35+18ln⁡hS = 35 + 18\ln h, by how much does SS increase each time hh doubles, from any starting value?

Sinusoidal models with a phase shift

In 2.5 the waves were in degrees. Now the angle is in radians, unless a degree sign is shown as in sin⁡x∘\sin x^{\circ}, and the wave can slide sideways.

The phase shift cc is a horizontal : c>0c > 0 moves the wave cc units right. With a>0a > 0, the wave crosses its principal axis going up at x=cx = c, and reaches its maximum a quarter of a period later.

0 4 8 12 16 20 24 12 16 20 24 28 32 Hours after midnight, t Temperature, T (°C) principal axis T = 22 min (4, 14) max (16, 30) phase shift c = 10
T(t)=8sin⁡(π12(t−10))+22T(t) = 8\sin\left(\frac{\pi}{12}(t - 10)\right) + 22 crosses T=22T = 22 going up at t=10t = 10.
Worked exampleThe temperature in a greenhouse, T ∘CT\,^{\circ}\mathrm{C}, tt hours after midnight is modelled by T(t)=8sin⁡(π12(t−10))+22T(t) = 8\sin\left(\frac{\pi}{12}(t - 10)\right) + 22. (a) Write down the amplitude and the phase shift. (b) Find the period. (c) Find when the greenhouse is warmest, and that temperature.

Try it first, then reveal one step at a time.

  1. (a) Amplitude 8 ∘C8\,^{\circ}\mathrm{C}; phase shift 1010 hours.
Worked exampleIn a city, the hours of daylight, DD, on day tt of the year are modelled by D(t)=asin⁡(b(t−c))+dD(t) = a\sin(b(t - c)) + d, with a>0a > 0, b>0b > 0 and 0<c<3650 < c < 365. The longest day has 15.715.7 hours of daylight, on day 171171, and the shortest has 8.18.1 hours. The period is 365365 days. (a) Find aa, bb, cc and dd. (b) Use your GDC to find for how many days of the year there are more than 1313 hours of daylight.

Try it first, then reveal one step at a time.

  1. (a) a=15.7−8.12=3.8a = \frac{15.7 - 8.1}{2} = 3.8 and d=15.7+8.12=11.9d = \frac{15.7 + 8.1}{2} = 11.9.
Check yourself
The height of a buoy above the sea floor, hh m, tt seconds after a timer starts is h(t)=0.4sin⁡(2.5(t−0.3))+11h(t) = 0.4\sin(2.5(t - 0.3)) + 11. Find the period, in seconds, to 3 s.f.
Check yourself
The depth of water at a dock, DD m, tt hours after 6 a.m. is D(t)=1.5sin⁡(0.5(t−2))+4D(t) = 1.5\sin(0.5(t - 2)) + 4, in radians. At what value of tt is the water first deepest after t=2t = 2?

Logistic models

goes on for ever, but real growth runs out of food or space. Rabbits on an island multiply fast, then the island fills up.

In the formula booklet
f(x)=L1+Ce−kx, L,k,C>0f(x) = \frac{L}{1 + C\mathrm{e}^{-kx}},\ L, k, C > 0

This logistic function is in the formula booklet. Use it where growth is restricted, such as a seedling's height.

At x=0x = 0, e0=1\mathrm{e}^{0} = 1, so f(0)=L1+Cf(0) = \frac{L}{1 + C}: the starting value. As xx grows, Ce−kxC\mathrm{e}^{-kx} shrinks towards 00, so f(x)f(x) gets closer and closer to LL.

0 2 4 6 8 10 12 0 200 400 600 800 1000 Years after release, t Number of rabbits, R carrying capacity R = 860 R(0) = 43 almost exponential growth slows
R(t)=8601+19e−0.62tR(t) = \frac{860}{1 + 19\mathrm{e}^{-0.62t}} never reaches 860860.
Worked exampleRabbits are released on an island. The number of rabbits tt years later is modelled by R(t)=8601+19e−0.62tR(t) = \frac{860}{1 + 19\mathrm{e}^{-0.62t}}. (a) Find the number released. (b) Write down the carrying capacity and interpret it. (c) Use your GDC to find when there are 600600 rabbits, to 3 s.f.

Try it first, then reveal one step at a time.

  1. (a) R(0)=8601+19=43R(0) = \frac{860}{1 + 19} = 43 rabbits.
Worked exampleThe height, hh cm, of a sunflower tt days after it is planted out is modelled by h(t)=2401+Ce−kth(t) = \frac{240}{1 + C\mathrm{e}^{-kt}}. It is 44 cm tall at t=0t = 0 and 102102 cm tall at t=30t = 30. (a) Find CC. (b) Find kk. (c) Find how many days it takes to reach 200200 cm, to 3 s.f.

Try it first, then reveal one step at a time.

  1. (a) h(0)=2401+C=4h(0) = \frac{240}{1 + C} = 4, so 1+C=601 + C = 60 and C=59C = 59.
Check yourself
The number of fish in a lake tt years after it is stocked is modelled by P(t)=54001+8e−0.3tP(t) = \frac{5400}{1 + 8\mathrm{e}^{-0.3t}}. Write down the carrying capacity.
Check yourself
The number of bacteria in a petri dish, NN thousand, tt hours after an experiment starts is N(t)=15001+499e−0.9tN(t) = \frac{1500}{1 + 499\mathrm{e}^{-0.9t}}. Find when the number reaches half the carrying capacity. Give your answer in hours, to 3 s.f.

Continuous piecewise models

If the rules of a piecewise model give different values where they join, the graph jumps. For a price or a concentration, a jump is often unrealistic.

You don't need a formal definition of continuity. To make a piecewise model continuous at x=px = p:

1. Substitute x=px = p into the rule on each side, even the one whose interval stops just before pp.

2. Set the two values equal.

3. Solve for the unknown parameter.

0 1 2 3 4 5 6 7 8 9 10 0 5 10 15 20 25 Mass of parcel, m (kg) Cost, C (dollars) a = 4: jump at 5 kg pieces meet at (5, 16) a = 8.5
With a=8.5a = 8.5 the pieces meet; with a=4a = 4 the price would jump down at 55 kg.
Worked exampleA courier charges CC dollars to send a parcel of mass mm kg, where C(m)=6+2mC(m) = 6 + 2m for 0<m≤50 < m \le 5 and C(m)=a+1.5mC(m) = a + 1.5m for 5<m≤105 < m \le 10. Find aa so that CC is continuous, and interpret the result.

Try it first, then reveal one step at a time.

  1. The pieces join at m=5m = 5.
Worked exampleA drug is given through a drip for 22 hours. Its concentration in the blood, cc mg per litre, tt hours after the drip starts is modelled by c(t)=6tc(t) = 6t for 0≤t≤20 \le t \le 2 and c(t)=Ae−0.35tc(t) = A\mathrm{e}^{-0.35t} for t>2t > 2. The model is continuous. (a) Find AA. (b) Find when the concentration falls back to 33 mg per litre, to 3 s.f.

Try it first, then reveal one step at a time.

  1. (a) Join at t=2t = 2: the first rule gives 6×2=126 \times 2 = 12, the second Ae−0.35×2=Ae−0.7A\mathrm{e}^{-0.35 \times 2} = A\mathrm{e}^{-0.7}.
Check yourself
A cyclist's distance from home, dd km, tt hours after setting off is d=18td = 18t for 0≤t≤1.50 \le t \le 1.5 and d=12t+qd = 12t + q for t>1.5t > 1.5. Find qq so that the model is continuous.
Check yourself
A company's monthly revenue, RR thousand dollars, in month tt is modelled by R=2t2R = 2t^{2} for 0≤t≤30 \le t \le 3 and R=k(t−1)R = k(t - 1) for t>3t > 3. Which value of kk makes the model continuous?

Models given in a question

A question may give a model you have never met. Use the same skills: substitute, use the GDC for graphs and equations, check large and small inputs, and interpret with units.

Worked exampleAfter an advert is shown, the number of visitors per hour to a website, VV, tt hours later is modelled by V(t)=900te−0.25tV(t) = 900t\mathrm{e}^{-0.25t}. (a) Find V(2)V(2) and interpret it. (b) Use your GDC to find the greatest number of visitors per hour, and when it happens. (c) Describe what the model predicts in the long term.

Try it first, then reveal one step at a time.

  1. (a) V(2)=900×2e−0.5=1091.7…V(2) = 900 \times 2\mathrm{e}^{-0.5} = 1091.7\ldots: about 10921092 visitors per hour, 22 hours after the advert.
Worked exampleA grower models the mass, mm kg, of a giant pumpkin tt weeks after it starts to swell by m(t)=36e−3.5e−0.6tm(t) = 36\mathrm{e}^{-3.5\mathrm{e}^{-0.6t}}. (a) Find the mass at t=0t = 0. (b) Find the mass the model predicts in the long term. (c) Find when the mass reaches 3030 kg, to 3 s.f.

Try it first, then reveal one step at a time.

  1. (a) m(0)=36e−3.5=1.087…m(0) = 36\mathrm{e}^{-3.5} = 1.087\ldots, about 1.091.09 kg.
Check yourself
The sound level, LL decibels, at a distance dd m from a loudspeaker is modelled by L=120−20log⁡10dL = 120 - 20\log_{10} d, d≥1d \ge 1. Find the sound level 88 m away, in decibels, to 3 s.f.
Check yourself
Total sales of a new game, SS thousand copies, tt months after launch are modelled by S(t)=50−35e−0.2tS(t) = 50 - 35\mathrm{e}^{-0.2t}. What does the value 5050 represent?

Common mistakes

  • Rounding kk or CC to 3 s.f. and then using it. Keep the GDC's full values, or later answers can be wrong in the last figure.
  • Calling the carrying capacity the starting value. The start is f(0)=L1+Cf(0) = \frac{L}{1 + C}; LL is the long-term limit.
  • Reading the phase shift before factorising. sin⁡(2t−6)=sin⁡(2(t−3))\sin(2t - 6) = \sin(2(t - 3)), so c=3c = 3, not 66.
  • Working in degrees on a radian model. Unless the question shows a degree sign, set the GDC to radians.
  • Using the wrong tt. Read what t=0t = 0 means: if tt is years after 2020, the year 2032 is t=12t = 12, not t=2032t = 2032.
  • Interpreting with no context. "LL is the asymptote" is not enough; say what it means, such as "the island can support at most about 860860 rabbits".

Exam tips

  • "Write down" the carrying capacity, or a phase shift from a model in the form asin⁡(b(x−c))+da\sin(b(x - c)) + d: read it from the model, with units. No working is needed.
  • "Find" a parameter from data: write the substituted equation before solving it on the GDC. It earns the method mark, and lets a wrong value earn follow-through later.
  • "Interpret": say what a value means in the situation, with units, such as "the long-term limit of the population".
  • "Comment" on a model: test it against the context, such as a negative yield for small inputs or growth that never stops.

Key takeaways

Learn thisExponential decay N0e−ktN_{0}\mathrm{e}^{-kt} with N0>0N_{0} > 0 and k>0k > 0: half-life =ln⁡2k= \frac{\ln 2}{k}, the same whatever the starting amount. After nn half-lives, (12)n\left(\tfrac{1}{2}\right)^{n} of it is left.
Learn thisf(x)=a+bln⁡xf(x) = a + b\ln x with x>0x > 0 and b≠0b \neq 0: a=f(1)a = f(1), and multiplying xx by mm adds bln⁡mb\ln m. No value at x=0x = 0 and no horizontal asymptote.
Learn thisf(x)=asin⁡(b(x−c))+df(x) = a\sin(b(x - c)) + d with b>0b > 0 and the angle in radians (assumed unless a degree sign is shown): amplitude ∣a∣|a|, period 2πb\frac{2\pi}{b}, principal axis y=dy = d, phase shift cc (to the right if c>0c > 0).
In bookletf(x)=L1+Ce−kx, L,k,C>0f(x) = \frac{L}{1 + C\mathrm{e}^{-kx}},\ L, k, C > 0
Learn thisLogistic L1+Ce−kx\frac{L}{1 + C\mathrm{e}^{-kx}} with L,C,k>0L, C, k > 0: value at x=0x = 0 is L1+C\frac{L}{1 + C}; carrying capacity LL, the horizontal asymptote y=Ly = L, approached but never reached.
Learn thisA piecewise model is continuous at a joining point x=px = p when both rules give the same value at pp: substitute pp into each rule, set them equal and solve for the parameter.

Practise this

7 past-paper questions on Logistic models in the AI HL bank, each with a mark scheme and a worked explanation.

Also: Logarithmic models (5) · Non-linear piecewise models (3)

Practise Logistic models →