Solving Optimization Problems

Every optimization problem has two equations hiding in it: the quantity you actually care about, like surface area or distance, and a constraint that ties its variables together, like a fixed volume or a point forced to sit on a curve. The constraint is what makes the problem solvable with one variable. You solve the constraint for one variable in terms of the other, substitute it into the quantity you're optimizing, and suddenly you have a single-variable function you can differentiate. Set that derivative to zero, solve, and check it actually gives a max or min, not a saddle point — use the first derivative test, second derivative test, or just check endpoints if the domain is closed.

This is why a soda can that minimizes metal for a fixed volume ends up tall and skinny instead of looking like the squat cans on shelves: real cans add extra material for seams and lids, but the pure math answer only cares about the constraint V = π r² h and the surface area formula. Once you collapse to one variable, say S(r) = 2 π r² + 2V/r, the calculus is identical to every related rates or curve sketching problem from earlier in the unit — you're just now choosing which critical point is the winner.

This topic is the payoff of the whole unit. Everything about increasing and decreasing intervals, concavity, and the extreme value theorem was building toward this moment: a word problem gets translated into a function, and the shape tools tell you where its extreme value lives. Distance problems often get minimized as squared distance instead, purely to avoid messy square root derivatives — a trick worth knowing because the next unit leans on the same substitution habit.