4.8 - Maximum and minimum problems

4.8 - Maximum and minimum problems

Use calculus to choose the largest or smallest value that a problem actually allows. You will turn a geometric constraint into one expression, minimise positive reciprocal expressions and check domain boundaries as well as stationary points.

Maximising an area with a constraint

You have 48 m48\,\mathrm{m} of fencing and a straight wall that provides one side of a rectangular enclosure. Would a square give the greatest area? The wall changes the problem: you pay for only three sides, so the best shape may surprise you.

You already know

An interior stationary point has zero first derivative. A negative second derivative there identifies a local maximum. Differentiate the quantity you want to optimise, then substitute into its original expression for the final value.

Let the two sides perpendicular to the wall each be xx metres, and the third fenced side be ll metres. The fencing constraint is 2x+l=482x+l=48, so l=48−2xl=48-2x. The area is A=xl=x(48−2x)A=xl=x(48-2x). Both dimensions must be positive, giving 0<x<240<x<24.

[DIAGRAM: asset_name: A rectangle against a wall; asset_slug: l2fm-fencing-constraint; caption: Only three sides use fencing, so the total fenced length is two short sides plus one long side; description: A rectangle lies against a horizontal fixed wall. The two perpendicular fenced sides are each labelled x metres; the opposite fenced side is labelled l metres. A note gives 2x plus l equals 48 and area A equals xl. The wall is clearly distinguished from the three sage fenced sides]
Only three sides use fencing, so the total fenced length is two short sides plus one long side

The constraint removes one variable. Once ll is written in terms of xx, you can use the differentiation you already know.

An interior optimum

For a differentiable quantity Q(x)Q(x), an interior maximum or minimum must satisfy

dQdx=0\frac{\mathrm{d}Q}{\mathrm{d}x}=0

Keep only solutions in the problem's domain, justify the nature and substitute into QQ for the requested value. Compare included boundary values when the domain has endpoints.
Not on the formulae sheet: learn it.

In the enclosure problem the area is a downward-opening quadratic, so its one stationary point will be the maximum on the allowed interval.

Worked example 1: three fenced sides

Question. Use calculus to work out the dimensions and maximum area of the rectangular enclosure made with 48 m48\,\mathrm{m} of fencing against a wall.

Plan. Differentiate the one-variable area, solve for xx and calculate the other dimension and area.

A=48x−2x2expand the areadAdx=48−4xdifferentiate48−4x=0⟹x=12stationary inputd2Adx2=−4<0maximuml=48−2(12)=24remaining fenced sideA=12×24=288original area formula\begin{aligned} A&=48x-2x^2 &&\text{expand the area}\\ \frac{\mathrm{d}A}{\mathrm{d}x}&=48-4x &&\text{differentiate}\\ 48-4x&=0\quad\Longrightarrow\quad x=12 &&\text{stationary input}\\ \frac{\mathrm{d}^2A}{\mathrm{d}x^2}&=-4<0 &&\text{maximum}\\ l&=48-2(12)=24 &&\text{remaining fenced side}\\ A&=12\times24=288 &&\text{original area formula} \end{aligned}

Why this step? xx measures a side length. The question asks for area too, so a final answer of just 1212 would leave the optimisation unfinished.

Answer. The dimensions are 12 m12\,\mathrm{m} by 24 m24\,\mathrm{m}, giving maximum area 288 m2288\,\mathrm{m}^2.

Check. 2(12)+24=482(12)+24=48, and 1212 lies in 0<x<240<x<24. The area tends to zero at either excluded end, so the interior maximum is also the greatest allowed area.

Worked example 2 changes only the amount of fencing to 64 m64\,\mathrm{m}. Build the constraint again; do not change the optimum by guesswork.

Worked example 2: a longer fence

Question. Use calculus to work out the dimensions and maximum area when the same kind of enclosure uses 64 m64\,\mathrm{m} of fencing.

Plan. Replace the total length in the constraint, then optimise the new area.

l=64−2x,0<x<32positive side lengthsA=x(64−2x)=64x−2x2one-variable areadAdx=64−4x=0⟹x=16stationarityd2Adx2=−4<0maximuml=64−32=32remaining sideA=16×32=512maximum area\begin{aligned} l&=64-2x,\qquad 0<x<32 &&\text{positive side lengths}\\ A&=x(64-2x)=64x-2x^2 &&\text{one-variable area}\\ \frac{\mathrm{d}A}{\mathrm{d}x}&=64-4x=0\quad\Longrightarrow\quad x=16 &&\text{stationarity}\\ \frac{\mathrm{d}^2A}{\mathrm{d}x^2}&=-4<0 &&\text{maximum}\\ l&=64-32=32 &&\text{remaining side}\\ A&=16\times32=512 &&\text{maximum area} \end{aligned}

Why this step? The two copies of xx both consume fencing, so the constraint contains 2x2x, not xx.

Answer. 16 m16\,\mathrm{m} by 32 m32\,\mathrm{m}, with maximum area 512 m2512\,\mathrm{m}^2.

Check. 16+16+32=6416+16+32=64. Completing the square gives A=512−2(x−16)2A=512-2(x-16)^2, confirming that no allowed area exceeds 512512.

The wall has a striking benefit. With the original 48 m48\,\mathrm{m}, fencing all four sides gives A=x(24−x)=144−(x−12)2A=x(24-x)=144-(x-12)^2, at most 144 m2144\,\mathrm{m}^2. Letting the wall provide one side doubles the best possible area to 288 m2288\,\mathrm{m}^2, even though the amount of fencing is unchanged.

Watch out

Wrong: x+l=48x+l=48. There are two fenced sides of length xx, so the correct constraint is 2x+l=482x+l=48. Count the actual sides before forming the expression.

The next enclosure changes the total again. Show the model before using calculus.

Minimising a positive reciprocal expression

In an expression containing both xx and 1/x1/x, increasing xx makes one positive term larger and the other smaller, so the smallest total balances the two effects.

The domain is essential. In the examples below x>0x>0: a negative stationary solution is not an alternative answer, and x=0x=0 is not defined at all.

Differentiating a reciprocal term

For constant bb and x≠0x\ne0,

bx=bx−1ddx(bx−1)=−bx−2=−bx2ddx(−bx−2)=2bx−3=2bx3\begin{aligned} \frac{b}{x}&=bx^{-1}\\ \frac{\mathrm{d}}{\mathrm{d}x}(bx^{-1})&=-bx^{-2}=-\frac{b}{x^2}\\ \frac{\mathrm{d}}{\mathrm{d}x}(-bx^{-2})&=2bx^{-3}=\frac{2b}{x^3} \end{aligned}

When b>0b>0 and x>0x>0, the second derivative of this term is positive.
Not on the formulae sheet: learn it.

Write the negative power before differentiating, then keep the positive-domain restriction beside your solution.

Worked example 3: an exact fractional input

Question. For x>0x>0, V=49x+81xV=49x+\frac{81}{x}. Use calculus to show that VV has a minimum and work out its minimum value.

Plan. Solve the zero-gradient equation, retain its positive solution and test it before evaluating VV.

V=49x+81x−1integer-power formdVdx=49−81x−2differentiate49−81x2=0⟹49x2=81multiply by x2x=97retain the positive rootd2Vdx2=162x3>0throughout x>0V(97)=49(97)+819/7=63+63=126evaluate the total\begin{aligned} V&=49x+81x^{-1} &&\text{integer-power form}\\ \frac{\mathrm{d}V}{\mathrm{d}x}&=49-81x^{-2} &&\text{differentiate}\\ 49-\frac{81}{x^2}&=0\quad\Longrightarrow\quad49x^2=81 &&\text{multiply by }x^2\\ x&=\frac97 &&\text{retain the positive root}\\ \frac{\mathrm{d}^2V}{\mathrm{d}x^2}&=\frac{162}{x^3}>0 &&\text{throughout }x>0\\ V\left(\frac97\right)&=49\left(\frac97\right)+\frac{81}{9/7}=63+63=126 &&\text{evaluate the total} \end{aligned}

Why this step? The equation for x2x^2 gives two algebraic roots, but only 9/79/7 is in the domain. A length-like positive input cannot be replaced by its negative root.

Answer. The minimum value is 126126, attained at x=97x=\frac97.

Check. For 0<x<9/70<x<9/7, 49−81/x2<049-81/x^2<0; for x>9/7x>9/7, it is positive. Thus VV falls to this value and rises afterwards throughout the positive domain.

Continue this lesson.

You’re previewing NovaLearn without an account. Create a free account to read the rest.

Create free account