Optimisation with constraints, boundaries and rejected roots
Introduit| English | Sens |
|---|---|
| feasible domain/ˈfiːzɪbl dəˈmeɪn/ | domaine réalisable |
Fermat's method of maxima, 1636
- In letters of about 1636, Pierre de Fermat described a way to find the largest or smallest value of a quantity: compare its value at $x$ and at a nearby $x + e$, and set the difference to zero. This was the first form of "set the derivative to zero". Packaging designers still use it: what size of corner square, cut from a sheet of card, makes the biggest open box?
- In optimisation, the answer lies at a stationary point or at an end of the feasible domain 可行域.
Building the function
- Name the variable, write the quantity to optimise in terms of it, and state the allowed values.
- Use the constraint (fixed area, fixed volume) to remove the second variable.
Squares of side $x$ are cut from a 30 cm by 20 cm card to make an open box. What is the domain of $x$?
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
The shorter side needs $20 - 2x > 0$.
Finding and testing the candidates
- Solve $f'(x) = 0$, keep only roots inside the domain, and test with $f''(x)$ or the sign of $f'$.
- On a closed interval, also check the endpoints.
Optimisation with constraints, boundaries and rejected roots
Express the quantity to optimise in one variable, with its feasible domain
Connect derivative calculations to their original point, domain and stated rate law.
A rectangular field against a wall uses 80 m of fence on three sides. Find the maximum area.
Difficulté 2/5 · 140 pièces + un petit bonus lors de la première réussite aujourd'hui
$A = x(80 - 2x)$; $A' = 0$ at $x = 20$.
Worked example: the open box
Squares of side $x$ cm are cut from the corners of a 24 cm square sheet, and the sides are folded up. Find the maximum volume.
- Model: $V = x(24 - 2x)^2$, with $0 < x < 12$.
- Differentiate: $V' = (24 - 2x)^2 - 4x(24 - 2x) = (24 - 2x)(24 - 6x) = 0$, so $x = 12$ or $x = 4$. Reject $x = 12$: it is outside the domain and gives $V = 0$.
- Test: $V'' = 24x - 192$, so $V''(4) = -96 < 0$: a maximum. $V = 4 \times 16^2 = 1024$ cm³.
Using $V = x(24 - 2x)^2$, find $V$ when $x = 3$.
Difficulté 2/5 · 140 pièces + un petit bonus lors de la première réussite aujourd'hui
$3 \times 18^2$; less than 1024.
In the open-box problem, $x = 12$ is rejected because it is outside $0 < x < 12$ and gives no box.
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
The sides would have length 0.
A can with fixed volume
- A closed can holds 330 cm³: $\pi r^2h = 330$, so $h = \dfrac{330}{\pi r^2}$ and the surface area is $S = 2\pi r^2 + \dfrac{660}{r}$.
- $S' = 4\pi r - \dfrac{660}{r^2} = 0$ gives $r = 3.74$ cm and $h = 7.49$ cm: the best can is as tall as it is wide.
For the can, $S = 2\pi r^2 + \dfrac{660}{r}$. Find $r$ that minimises $S$, to 3 s.f.
Difficulté 2/5 · 140 pièces + un petit bonus lors de la première réussite aujourd'hui
$r^3 = \tfrac{660}{4\pi}$.
Endpoints matter
- If a fence must be between 10 m and 30 m, and the stationary point is at 40 m, the best value is at an endpoint.
- Compare $f$ at each endpoint and each valid stationary point.
$f(x) = x^3 - 12x$ on $0 \le x \le 3$. Find the maximum value of $f$.
Difficulté 2/5 · 140 pièces + un petit bonus lors de la première réussite aujourd'hui
Stationary at $x = 2$ (a minimum, $-16$); endpoints $f(0) = 0$, $f(3) = -9$.
Not every root of $f'(x) = 0$ is the answer. Reject roots outside the domain, check whether each one is a maximum or minimum, and compare with the endpoints before you decide.
Write the quantity in one variable using the constraint, state its domain, solve $f'(x) = 0$, reject roots outside the domain, classify with $f''$, and compare with the endpoints. Give the answer in context, with units.
A window is a rectangle topped by a semicircle, with a perimeter of 6 m. Find the width that lets in the most light, to 3 s.f.
Constraint, derivative, test.