Last week, we discussed optimization problems where all of the numbers were given. For example, we solved problems where we were given that the surface area of, say, a box was 100cm2 and we had to determine the maximum possible volume of the box. This week, we will be considering problems like the following: Example 1. An open-topped box of surface area A is going to be built with a square base. Which dimensions should the box have to maximize the volume? This is similar to a problem we’ve encountered before. However, the number 100cm2 was replaced by the parameter A. The secret to this problem is that it’s not any different from the problem where A was replaced by a number. Therefore, to solve the problem, we treat A as if it were any old number, e.g. a constant. Our answer will depend on the area A. Let’s start by writing down the equations for the volume and the surface area. If s is the side length of the base, and h is the height, we have V = s2 h and A = 4sh + s2 . As usual, we want to get V in terms of a single variable. In this case, it makes sense to get v in terms of s, because h is easier to solve for in the equation for A. So we solve for h: A − s2 h= 4s and we get A s 2 V =s − 4s 4 So As s3 − . 4 4 On the face of it, it looks like v depends on both A and s, and in fact it does. However, for this problem A is a parameter - it’s fixed and we can’t change it. In other words, it should be treated as if it were a constant number, like 100cm2 . So when we differentiate, we treat A like a constant. We will now differentiate: V = V′ = A 3s2 − 4 4 setting V ′ = 0 gives: 3s2 = A so s= r 1 A . 3 Note that this gives a volume of A V = 4 1/2 3/2 A 1 A − 3 4 3 There is a power of A3/2 in both terms so we factor it out: 1 1 √ − √ V = A3/2 4 3 12 3 which is 3/2 V =A 2 √ 12 3 , a positive number. Finally, we check √ the endpoints: s can be as small as 0, giving a volume of 0, or as large as A, giving a volume of A3/2 A3/2 − 4 4 q which is zero. So we should select s to be A 3. V = Notice that the difficulty here is in determining which of the critical points and endpoints gives the largest q value. Here, it took some work to see that the largest value you got was at A 3 - it wasn’t even immediately obvious that this number was positive! Sometimes, you’ll have a problem with more than one parameter in it: Example 2. You have a piece of wire of length ℓ, and you are going to cut it. You will bend one piece of the wire to make a regular polygon with n sides, and bend the other piece of the wire to make a circle. How should you cut the wire in order to maximize the area enclosed by both the polygon and the circle? The area of a regular polygon with n sides is Apolygon s2 n , 4 tan nπ where s is the side length of the polygon. We compute the total area A = Apolygon + Acircle and total perimeter ℓ = ℓpolygon + ℓcircle the perimeter of the polygon is ℓpolygon = ns 2 and the circumference of the circle is ℓ − ns so the radius is ℓ − ns 2π and the area is π ℓ − ns 2π 2 . Thus, we get the total area in terms of s: s2 n +π A= 4 tan nπ ℓ − ns 2π 2 Notice that A and s are the only variables here: n and ℓ are parameters. We’ll expand the square so that it’s easier to deal with later: ℓ2 − 2ℓns + n2 s2 s2 n + π 4π 2 4 tan πn A= which is A= s2 n ℓ2 − 2ℓns + n2 s2 + 4π 4 tan πn and now we differentiate. Remember s is the only variable here; we treat ℓ and n as constants: 2sn2 − 2ℓn 2sn + A′ (s) = 4π 4 tan nπ we set A′ = 0: 0= 2sn 2sn2 − 2ℓn π + 4π 4 tan n we can now divide by 2n since it occurs in every term: 0= s 4 tan π n This is a linear equation in s: 1 4 tan ℓ = 4π so we get: + sn − ℓ 4π n π + 4π n ! ℓ s= 4π 1 π 4 tan( n ) 3 + n 4π s which comes out to s= ℓ π π tan( n ) +n . With this value of s, we get that we should cut the wire at length sn, which is sn = ℓ π π n tan( n ) +1 . I’ll leave it to you guys to determine whether or not this is the maximum. 4