Introduction to Optimization Problems

A farmer with a fixed length of fence gets more area from a square field than a long skinny rectangle, because area depends on how you split the perimeter between length and width, not just how much fence you use. Optimization problems formalize this: some quantity (area, volume, cost) depends on two or more variables linked by a constraint. Use the constraint to eliminate one variable, and you are left with a single function of a single variable — exactly the kind of function you already know how to maximize with derivatives.

The method: name your variables, write down the quantity you want to optimize, write down the constraint, solve the constraint for one variable, and substitute. That gives you one function on some domain. Take its derivative, set it to zero, check critical points and endpoints — the same candidates test from finding absolute extrema earlier in the unit. Nothing new happens at the calculus step; the new work is translating words into that one function.

The shape of the constraint controls the answer, which is why a square is not always optimal. With all four sides fenced, perimeter 2x+2y=P forces the max-area rectangle to be a square. But drop one side — say a river needs no fence — and the constraint becomes 2x+y=P instead. That changes which derivative equals zero, and the optimal rectangle comes out twice as long as it is wide. Same method, different constraint, different answer.