5.7 · SLPaper 1Paper 2

Optimisation problems

In a minute

means finding the best value of something in a real situation: the greatest profit, the least cost, the largest volume. Write the quantity as a function of one variable, solve where its is zero, then check that point against the ends of the allowed values. Always finish by answering the question in context.

You'll be able to

  • Write the quantity to optimise as a function of one variable, using a constraint to remove a second unknown.
  • Solve f′(x)=0f'(x) = 0 by hand or with your GDC, and justify whether the point is a maximum or a minimum.
  • Find the greatest or least value on a restricted domain, checking the endpoints and whole-number values.
  • Interpret an optimum in context, such as profit, cost, area or volume, and know that motion (kinematics) questions are not set at SL.
Before you start3 quick questions on what this builds on
Check yourself
Differentiate f(x)=3x2+50xf(x) = 3x^{2} + \frac{50}{x}.
Check yourself
Find the xx-coordinates of the of y=x3−27xy = x^{3} - 27x.
Check yourself
On the domain 0≤x≤100 \le x \le 10, a function has f(0)=5f(0) = 5, its only at (2,12)(2, 12), and f(10)=30f(10) = 30. What is the greatest value of ff on this domain?

Revise first: 2.4 Key features of graphs, 5.1 Limits and the derivative as a rate of change, 5.2 Increasing and decreasing functions, 5.3 Differentiation of integer-power polynomials and 5.6 Local maximum and minimum points

Finding the best value

Which price brings in the most money? Which box uses the least card? Finding the greatest or least value of a quantity is . On a smooth graph the best value is often at a , where the derivative is zero.

1. Write the quantity you want to optimise as a function of one variable. Note the values that variable can take, its .

2. Differentiate. Then solve f′(x)=0f'(x) = 0, by hand or with your GDC.

3. Justify that the point is the maximum or minimum you want. Check the ends of the domain too.

4. Answer the question asked, in context and with units. If it asks for the greatest profit, give the profit, not just xx.

To justify at SL, show that f′(x)f'(x) changes from positive to negative (a ) or negative to positive (a ). Or say which way the parabola opens, or sketch the GDC graph with the point marked.

Worked exampleA workshop refurbishes and sells xx bicycles a month. Its monthly profit, in dollars, is modelled by P(x)=−1.5x2+168x−2400P(x) = -1.5x^{2} + 168x - 2400, for 0≤x≤1000 \le x \le 100. (a) Find P′(x)P'(x). (b) Find the number of bicycles that gives the greatest monthly profit, and justify that it is a maximum. (c) Find the greatest monthly profit.

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

  1. (a) Use the power rule on each term: P′(x)=−3x+168P'(x) = -3x + 168.

If you can't differentiate the model by hand, as with ex\mathrm{e}^{x} at SL, graph it on your GDC and use the maximum or minimum tool. Write down the function you graphed: that line earns the method mark.

Worked exampleA streaming service charges xx dollars a month. It models the number of subscribers as n=5000e−0.08xn = 5000\mathrm{e}^{-0.08x}. (a) Write down an expression for the monthly revenue RR, in dollars, in terms of xx. (b) Use your GDC to find the price that gives the greatest revenue, and the greatest revenue. Give the revenue to the nearest cent. (c) Interpret the shape of the model.

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

  1. (a) Revenue is price times number of subscribers: R=5000xe−0.08xR = 5000x\mathrm{e}^{-0.08x}.
Check yourself
The yield of a tomato plant, in kg, is modelled by Y=1.2+0.08f−0.0004f2Y = 1.2 + 0.08f - 0.0004f^{2}, where ff is the mass of fertiliser in grams, 0≤f≤2000 \le f \le 200. Find the greatest yield, in kg.
Check yourself
For a profit model P(x)=−2x2+40x−75P(x) = -2x^{2} + 40x - 75, a student solves P′(x)=0P'(x) = 0, gets x=10x = 10 and writes "maximum". Which reason would earn the reasoning mark?

At SL, exam optimisation questions use contexts such as profit, cost, area and volume. Questions on motion (kinematics) are not set at SL.

Using a constraint

Often the quantity depends on two unknowns, such as a length and a width. A fixed fact links them: an area, a volume or an amount of material. This is the constraint. Rearrange it for one unknown and substitute, so one variable is left.

Rectangular pen against a barn wall, fenced on three sides barn wall (no fence) x m y m y m area 450 m² fence
Fence on three sides only: the length of fence is x+2yx + 2y.
Worked exampleA farmer builds a rectangular pen of area 450 m2450\text{ m}^{2} against a long barn wall. Fencing is needed on the other three sides only. The side parallel to the wall is xx m and the other two sides are yy m each. (a) Show that the length of fencing is L=x+900xL = x + \dfrac{900}{x}. (b) Find the dimensions that use the least fencing, and justify that it is a minimum. (c) Find the least length of fencing.

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

  1. (a) The constraint is the area: xy=450xy = 450, so y=450xy = \dfrac{450}{x}.

A classic design problem fixes the material and asks for the largest volume. The area and volume formulas are in the prior-learning section of the formula booklet.

Worked exampleA closed cylindrical water tank has radius rr m and height hh m. It is made from 24 m224\text{ m}^{2} of sheet metal, which forms the whole surface. (a) Show that h=12πr−rh = \dfrac{12}{\pi r} - r. (b) Show that the volume is V=12r−πr3V = 12r - \pi r^{3}. (c) Find the radius that gives the greatest volume. (d) Find the greatest volume. Give your answers to 3 s.f. (e) Comment on the shape of this tank.

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

  1. (a) The surface is two circles plus the curved side: 2πr2+2πrh=242\pi r^{2} + 2\pi rh = 24. So 2πrh=24−2πr22\pi rh = 24 - 2\pi r^{2} and h=24−2πr22πr=12πr−rh = \dfrac{24 - 2\pi r^{2}}{2\pi r} = \dfrac{12}{\pi r} - r.
Check yourself
A closed storage crate has a square base of side xx cm and a volume of 2000 cm32000\text{ cm}^{3}. Which expression gives its total surface area, S cm2S\text{ cm}^{2}, in terms of xx?
Check yourself
For the crate above, S=2x2+8000xS = 2x^{2} + \frac{8000}{x}. Find the value of xx that makes the surface area least. Give your answer in cm to 3 s.f.

Endpoints, whole numbers and the answer

On a restricted domain the greatest or least value can be at an end, not at a stationary point. So work out the quantity at every stationary point inside the domain and at both endpoints. The greatest is the .

Membership model on 0 to 15 months: the endpoint is higher than the local maximum 51015102030 t N local max (4.73, 26.3) endpoint (15, 28.75)
The local maximum is not the greatest value: the end of the domain, t=15t = 15, is higher.
Worked exampleThe number of members of a new climbing gym, in hundreds, is modelled by N(t)=0.05t3−1.2t2+8t+10N(t) = 0.05t^{3} - 1.2t^{2} + 8t + 10, where tt is the number of months after opening, 0≤t≤150 \le t \le 15. (a) Find N′(t)N'(t). (b) Use your GDC to solve N′(t)=0N'(t) = 0. (c) Find the greatest number of members during the first 15 months. (d) Comment on using this model after 15 months.

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

  1. (a) N′(t)=0.15t2−2.4t+8N'(t) = 0.15t^{2} - 2.4t + 8.

Some quantities must be whole numbers: chairs, tickets, workers. If the stationary point is not a whole number, work out the quantity at the whole numbers on each side and compare them.

Worked exampleA carpenter makes nn chairs a week. The average cost per chair, in dollars, is modelled by A(n)=0.7n+30+1800nA(n) = 0.7n + 30 + \dfrac{1800}{n}, for n>0n > 0. (a) Find A′(n)A'(n). (b) Find the number of chairs a week that gives the least average cost, and that cost.

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

  1. (a) Write 1800n\dfrac{1800}{n} as 1800n−11800n^{-1}: A′(n)=0.7−1800n2A'(n) = 0.7 - \dfrac{1800}{n^{2}}.
Check yourself
A bakery's daily profit, in dollars, from baking xx dozen cookies is P(x)=−2x2+64x−150P(x) = -2x^{2} + 64x - 150. Find the greatest daily profit, in dollars.
Check yourself
The price of a share, in dollars, is modelled by S(t)=0.4t3−4.8t2+14.4t+12S(t) = 0.4t^{3} - 4.8t^{2} + 14.4t + 12, where tt is the time in months, 0≤t≤90 \le t \le 9. Find the greatest share price in this period, in dollars to the nearest cent.

Common mistakes

  • Giving where the optimum happens when the question asks for the optimum itself. For "the greatest profit", substitute back into P(x)P(x) and give the profit.
  • Writing "maximum" with no reason. Show f′(x)f'(x) changing sign, say a quadratic model's parabola opens downwards, or sketch the GDC graph with the point marked.
  • Stopping at the local maximum on a restricted domain. Work out the quantity at both endpoints too: the greatest value may be at an end.
  • Keeping a value the situation rules out, such as a negative length, or giving a fraction of a chair. Reject it and say why, or compare the whole numbers either side.
  • Rounding an intermediate value, such as the radius, before using it. Keep the GDC's full value and round only the final answer.

Exam tips

  • "Find the value of xx that maximises ...": write P′(x)=0P'(x) = 0 or the function you graph before the GDC answer. That written set-up earns the method mark even if the final value slips.
  • "Justify" a maximum or minimum: give a reason, such as the sign of f′(x)f'(x) on each side or "negative quadratic, so its only stationary point is a maximum". A bare "maximum" earns no reasoning mark.
  • In a multi-part optimisation question, use the given "show that" result in later parts even if you couldn't show it. A wrong earlier answer used correctly can still earn follow-through (FT) marks.

Key takeaways

Learn thisTo optimise: write the quantity as a function of one variable, solve f′(x)=0f'(x) = 0, justify the maximum or minimum, compare it with the endpoints of the domain, then answer in context with units.
Learn thisWith two unknowns, use the constraint (a fixed area, volume or amount of material) to write one in terms of the other before you differentiate.
Learn thisAt a stationary point, f′(x)f'(x) changing from positive to negative means a local maximum; from negative to positive means a local minimum.
Learn thisFor a function you can differentiate on all of a≤x≤ba \le x \le b, the greatest or least value is at a stationary point inside the domain or at an endpoint, so compare them all.
Learn thisIf the variable must be a whole number and the optimum is at a stationary point that is not, compare the whole numbers on each side of it.
Learn thisWith a GDC, write down the function you graph or the equation you solve, keep full values, and round only the final answer.

Practise this

6 past-paper questions on Modelling with differentiation in the AI HL bank, each with a mark scheme and a worked explanation.

Practise Modelling with differentiation →