Sometimes even the simplest problems can lead to interesting places. I got interested in this ordinary textbook problem a few months ago while observing an AB calculus class. Here is how my thinking went written as a series of exercises:

Original exercise: Find the point(s) on the parabola *y = x*^{2} that are closest to the point (0, 2).

The solution is straightforward:

Let z*(*x*)* be the distance from (0, 2) to the point (*x*, *x*^{2}). Then and setting the derivative equal to zero, we find the critical points at *x* = 0 a local maximum and and both absolute minimums located symmetrically to the *y*-axis.

The students had some trouble understanding what was going on. I suggested they consider a point moving along the parabola starting high on the left side. As the point comes down the graph its distance from gets shorter until it reaches some minimum length somewhere. Then the distance gets longer again until – where? Obviously, at the origin. Then symmetry takes over and the distance decreases again until it reaches a second minimum directly across the *y-*axis from the first point, after which it increases forever.

Exercise 1: Sketch the graph of the distance function, *z*(*x*), on top of the graph of the parabola. Give the exact coordinates of the maximum point *without* doing any computations. Discuss the *y*-coordinates of the minimum points.

It then occurred to me that there is nothing sacred about (0, 2). We could use any point on the *y*-axis, . Or could we? I pictured a circle centered at the point (0, *a*) on the *y*-axis tangent to the parabola (i.e. tangent to the tangent line of the parabola). The radius of such a circle is the minimum distance for that *y*-axis center point.

Exercise 2: If the point is below the origin then obviously the minimum distance is the distance directly up to the origin. There are places above the *x*-axis where the origin is the closest point. For what values of *a* is the origin the closest point? Use the circle mentioned above to explain how this is possible.

The distance graph that we drew above has a W shape similar to the shape of some 4^{th} degree polynomials (note: *z*(*x*) is not a 4^{th} degree polynomial, it is not even a polynomial). So next I drew the graph of *z’*(*x*), the first derivative of *z*; this graph is similar to a cubic polynomial (again it is not a polynomial)

Exercise 3: Without actually computing the derivative’s equation, sketch the graph of the derivative with the graph of the parabola and the distance function’s graph. Be sure the critical points are where they should be.

Exercise 4: Use a graphing program or a graphing calculator do draw the parabola, *z*(*x*) and *z’*(*x*) and use it to check your sketches.

Project: Using a suitable program (Winplot, Geogebra, etc.) set up an animation that will show all the features discussed above. It should show the parabola which does not move, the movable point (0, *a*), the circle tangent to the parabola, *z*(*x*) and *z’*(*x*). These last 3 should move with the slider for *a*.

At this point I noticed that the derivative changed concavity 4 times (up-down-up-down). The origin is one of 3 points of inflection. I wondered if there was anything interesting about the location of the points of inflection.

Exercise 5: Find the coordinates of the points of inflection of *z’*(*x*). Are they related to any of the features in the problem? (Use a CAS for this one.)

The solutions can be found here: Closest Point Problem Solution.