Extrema, inflexion and optimisation
In a minute
A or sits where ; the sign of or says which. The same idea finds the largest area or volume in a real problem. The sign of also shows which way a graph bends, and it switches at a .
You'll be able to
- Find the coordinates of and points by solving .
- Tell a maximum from a minimum by the sign change of or the sign of .
- Solve optimisation problems about area, volume and profit.
- Say where a graph is or from the sign of .
- Find , with zero or non-zero gradient, and know why alone is not enough.
Before you start2 quick questions on what this builds on
Revise first: 5.2 Increasing and decreasing functions and 5.7 Second derivative and concavity
Local maximum and minimum points
Follow the graph of from left to right. It climbs to a hilltop at , drops into a valley at , then climbs again.
At a hilltop or a valley of a smooth curve the tangent is horizontal, so . So local maxima and minima are , and you find them by solving .
Move the slider to slide the point along . The sign of flips from to at the maximum, , and from to at the minimum, .
f(x) = x³ − 3x, so f′(x) = 3x² − 3
f′(−1.4) =
1. Find and solve . The solutions are the -coordinates of the stationary points.
2. Substitute each solution into , not , to get its -coordinate.
3. Test one value of on each side of every stationary point. to is a local maximum, to a local minimum.
Try it first, then reveal one step at a time.
- Differentiate: .
- Set and factorise: , so or .
- Substitute into for the -coordinates: and .
- Test one value in each gap: , and .
- At , goes from to , so is a local maximum. At it goes from to , so is a local minimum.
- Local maximum , local minimum
A factor that is always positive, such as , never changes the sign of .
Try it first, then reveal one step at a time.
- Use the product rule from 5.6: .
- Take out the common factor: .
- Solve . Since for every , the factor is never . So or .
- Find the -coordinates: and .
- The sign of is the sign of : positive for , negative for and positive for .
- So is a local maximum ( to ) and is a local minimum ( to ).
- Local maximum , local minimum
The second derivative test
Near a local minimum the curve bends like a cup. Its gradient rises from negative to positive, so the is positive: (see 5.7).
Near a local maximum it bends like an upside-down cup, so the gradient falls and . Checking one sign is often quicker than a sign table.
Try it first, then reveal one step at a time.
- Write as , then differentiate: .
- Solve : , so and . The domain is , so reject . That leaves .
- Find the -coordinate: .
- Differentiate again: .
- Substitute: , so is a local minimum.
- Local minimum
Try it first, then reveal one step at a time.
- Differentiate: .
- Solve : . For this gives or .
- Differentiate again: .
- Test each point: , so is a local maximum. , so the other point is a local minimum.
- Find the -coordinate of the maximum: .
The test fails when . Both and have .
Yet the gradient of , , goes from to , so the origin is a minimum. The gradient of , , stays positive, so there it is neither. So go back to the sign of .
Optimisation
Optimisation means finding the largest or smallest value of something, such as an area, a volume or a profit. That value is usually a local maximum or minimum.
A problem often has two unknowns linked by a fact, called a constraint. Use the constraint to write the quantity in one variable, then differentiate.
1. Name the variables and write the quantity, say , in terms of them. Use the constraint to leave one variable, and note the values it can take.
2. Solve , rejecting impossible values such as a negative length. Justify the maximum or minimum.
3. Answer the question asked. If it asks for the largest area, give , not just .
Try it first, then reveal one step at a time.
- (a) The surface is the base plus four sides: , so .
- Substitute into : , as required.
- (b) Differentiate and set to zero: , so . A length is positive, so and not .
- Justify: , which is at . So this is a maximum.
- Find the volume: . Check the height makes sense: cm.
On Paper 2, may be too hard to solve by hand. Graph the function and use your GDC's maximum or minimum tool.
Try it first, then reveal one step at a time.
- The model uses radians, so set your GDC to radian mode first.
- Write down what you graph: for .
- Use the maximum tool. It gives and .
- Check the ends of the interval: and . Both are lower, so the largest value is at .
- Round only at the end: the maximum temperature is degrees Celsius.
Concavity and points of inflexion
In the figure in the first section, the curve bends like an upside-down cup left of and like a cup right of it. Where the bend switches is a .
The gradient at a point of inflexion may be zero, as for at the origin, where the tangent is horizontal. Or it may not: on the cubic in the figure, .
1. Solve , then check that changes sign there with a value on each side.
2. Find the -coordinate from . For the gradient there, work out .
Try it first, then reveal one step at a time.
- Differentiate twice: and .
- Solve : , so .
- Check the sign change: and . So the graph is concave-down for and concave-up for .
- Find the -coordinate: .
- The gradient there is , which is not zero: the curve is still falling as it passes through.
Try it first, then reveal one step at a time.
- Differentiate twice: and .
- Solve : or .
- Check the signs: , and . The sign changes at both, so both are points of inflexion.
- Find the -coordinates: and .
- Find the gradients: , so has zero gradient. , so has a non-zero gradient.
Common mistakes
- Putting the -value into for the -coordinate. That gives ; use .
- Calling a point an inflexion just because . Check that changes sign: has a minimum at the origin.
- Using for concavity. Concave-up means ; solving finds where is increasing instead.
- Keeping a value the problem rules out, such as a negative length. Reject it and say why.
- Giving where the maximum happens when the question asks for the maximum value. Substitute back and give the value.
Exam tips
- "Justify" a maximum or minimum: give with its sign, such as , or the sign of on each side. A bare "minimum" is not a justification.
- On Paper 2, write down the function you graph, then give the GDC's maximum or minimum to 3 s.f. Check the ends of the domain too.
- For where a graph is concave-up, give strict inequalities from , such as .
- In an optimisation "show that", work forwards from the constraint to the given formula. Use that formula in later parts even if you couldn't show it.
Key takeaways
Practise this
17 past-paper questions on 5.8 in the AA HL bank, each with a mark scheme and a worked explanation.