Global maxima and minima and optimisation on restricted domains: the closed interval method, endpoints and discontinuities as candidates, domain restrictions in modelling problems and writing a contextual conclusion (local and global extrema and endpoint checks carry over from 2017; formal definitions and discontinuity checks are added in 2024)
“Distinguish local from global maximum and minimum values, find global extrema of a function on a given domain by comparing stationary points, endpoints and discontinuities, and justify an optimisation answer against the domain of the model”
The global maximum or minimum of a function on a domain is found by comparing at stationary points inside the domain, at the endpoints and at any discontinuity or point of non-differentiability. Reject candidates outside the domain, justify your choice, and write a conclusion in context with units. A local extremum is only extreme near itself. The 2017 course already distinguished local from global extrema and checked endpoints; the 2024 Mathematics Advanced syllabus adds formal definitions of global extrema and names discontinuities of and as places to check.
What this dot point is asking
Optimisation is not new to Mathematics Advanced: the 2017 course had maximum and minimum problems. The 2017 course already distinguished local from global extrema and checked endpoints. The Mathematics Advanced 11-12 Syllabus (2024) adds more precise language and a more careful method. Under Applications of calculus, NESA gives formal definitions of a global maximum and minimum, names discontinuities of and as places to check alongside the endpoints of the domain, and asks you to model a real situation with a function with an appropriate domain restriction, then interpret the answer in context.
So an optimisation answer has three stages: build the function and its domain, find all the candidates, and write a conclusion that respects the domain.
Think of walking over a hilly path that starts at one gate and ends at another. The highest point you stand on might be a hilltop on the way, but it might simply be the gate you finish at, if the path climbs steadily all the way. A local peak is only higher than its neighbours. The global highest point is the highest anywhere on the path, so you must compare the hilltops with both gates.
The answer
- Global (absolute) maximum of on a domain: the greatest value takes on that domain. Global minimum: the least value.
- Local maximum or minimum: greatest or least in a small neighbourhood only.
- On a closed interval , compare at: the stationary points in , the endpoints and , and any points where is discontinuous or not differentiable. The largest value is the global maximum and the smallest is the global minimum.
The closed interval method
For a continuous function on :
- Differentiate and solve . Keep only the solutions inside .
- List points in the domain where does not exist (see differentiability).
- Evaluate at every candidate and at both endpoints.
- State the largest and smallest values and where they occur.
A continuous function on a closed interval always has a global maximum and a global minimum. If the interval is open, or the function has a break, one of them may fail to exist. For example on has no global maximum or minimum, since and are never reached.
Why endpoints matter
An endpoint of the domain is a boundary where the function simply stops, so it can have the greatest or least value even though the gradient there is not zero. In the first question above, on has its global maximum at the endpoint , not at the turning point.
Discontinuities
If has a jump, check the value at the jump and the values approached on each side. A global extremum must be a value the function actually takes. When one side approaches a number without reaching it, that number is not attained.
Modelling with a restricted domain
In a practical problem the variable is limited by the situation: lengths are positive, a wall may be a fixed length, a box cannot have a negative base. Write down the domain when you write the function. Then:
- Use calculus on the function to find stationary points.
- Reject any stationary point that falls outside the domain, with a reason.
- Compare the remaining stationary points and the domain endpoints.
- Finish with a sentence in context, with units.
Maximise a product with a constraint
Two positive numbers have a sum of . Find the maximum value of their product.
Build the model
Let the numbers be and with . The product is .
Differentiate and test
gives , and , so it is a maximum. . The endpoints give .
Conclude
The maximum product is , when both numbers equal .
Minimise a cost on a restricted domain
The cost, in dollars, of running a machine for hours is for . Find the minimum cost.
Locate the vertex
at , which is outside .
Check the endpoints
and . On the domain , so is decreasing and the minimum is at .
Conclude
The minimum cost is $7, at hours.
- Forgetting the endpoints
- A stationary point is not automatically the global maximum or minimum.
- Keeping a stationary point outside the domain
- If the algebra gives but the domain is , the point is rejected, and the answer is at an endpoint.
- Calling a local extremum global
- Compare values at all candidates before you use the word "greatest".
- Missing the domain
- Write the restriction when you build the function, or you lose the mark for justification.
- No conclusion in context
- A bare "" is incomplete. State "the maximum volume is cm when the cut is cm".
Make a small two-column table of candidate values and the matching values. It makes the comparison visible to the marker and stops you skipping an endpoint. In a "show that" model question, show the substitution that eliminates the second variable. Finish every modelling answer with a sentence, including units.
Exam-style questions
Questions in the style of NESA exam questions on this dot point, each with a worked answer. They are written by ExamExplained unless tagged "Past paper"; the year shows the paper a question is modelled on.
HSC-style4 marksA closed rectangular box has a square base of side cm and a total surface area of cm. Show that its volume is and find the maximum volume.
Show worked answer →
Let the height be . Surface area , so . Then .
gives , so (the positive root). , so this is a maximum. cm.
Markers award one mark for the height in terms of , one for , one for with a reason for the root chosen and one for the volume with the nature established.
HSC-style3 marksFind the global maximum and minimum of on .
Show worked answer →
, so and . Then and . Endpoints: and .
The global maximum is at and the global minimum is at . The local maximum at and the local minimum at are neither. Markers reward the full list of candidates and the correct comparison.
Practice questions
Original practice questions graded from foundation to exam level, each with a full worked solution. Try them before revealing the solution.
foundation3 marksFind the global maximum and global minimum values of on the interval .
Show worked solution →
- Stationary point
- gives , which lies in the interval, and .
- Endpoints
- and .
- Compare
- The values are , and . The global maximum is (at ) and the global minimum is (at ).
Marker's note: one mark for the stationary point, one for both endpoint values and one for the correct comparison.
core3 marksFind the global maximum and global minimum of on .
Show worked solution →
Stationary points. gives . Then and .
Endpoints. and .
Compare , , and . The global maximum is at . The global minimum is , reached at both and .
Marker's note: the local maximum is not the global maximum. The function takes the same minimum value at two different -values.
core3 marksFind the global maximum and minimum of on .
Show worked solution →
- Stationary point
- gives , so (the root is outside the interval). .
- Endpoints
- and .
- Compare
- The global minimum is at . The global maximum is , and it occurs at both endpoints and .
Marker's note: one mark for the derivative and root inside the domain, one for both endpoints, one for the conclusion.
exam5 marksA square sheet of card of side cm has a square of side cm cut from each corner, and the sides are folded up to make an open box. (a) Show that the volume is and state the domain. (b) Find the value of that gives the maximum volume, and that volume.
Show worked solution →
(a) Build the model. The base is a square of side and the height is , so . The base side must be positive and must be positive, so the domain is ; the closed interval is used when checking endpoints, where the volume is .
(b) Differentiate. Expanding, , so
Stationary points are and . Only is inside the open domain.
Evaluate. . At the endpoints and . So the global maximum volume is cm when cm.
Marker's note: one mark for the model, one for the domain, one for the derivative, one for rejecting with a reason and one for the volume with units.
exam4 marksThe function is defined on by for and for . Find the global maximum and global minimum values of and say where each occurs.
Show worked solution →
Examine each piece. On , is increasing, from and approaching as from the left. The value is not attained by this piece.
On , is decreasing, from to .
Compare. The function is discontinuous at (the left piece approaches but ). The largest value actually reached is , so the global maximum is at . The smallest value is , the global minimum at .
Marker's note: one mark for each piece analysed, one for recognising the jump at and one for the correct extrema. There are no stationary points to find here.
exam5 marksA farmer has m of fencing to enclose a rectangular paddock using an existing straight wall as one side, so only three sides need fencing. The wall is m long, so the side parallel to it can be at most m. Let metres be the length of each of the two sides perpendicular to the wall. (a) Show that the area is and state the domain of . (b) Find the maximum area and the value of that gives it.
Show worked solution →
- (a) Model
- The side parallel to the wall is , so . The parallel side must satisfy , giving , and it must be non-negative, giving . So the domain is .
- (b) Differentiate
- , which is zero at . But is outside the domain, so there is no stationary point to consider.
- Check the endpoints
- On , , so is decreasing and the maximum is at the left endpoint:
The maximum area is m when m (so the parallel side is m).
Marker's note: one mark for the model, one for the domain, one for rejecting , one for the endpoint comparison and one for the contextual conclusion. Without the wall limit, the unrestricted maximum would be m, which is not allowed here.