Thursday, March 2, 2017

The Inscribed Square Problem

This week’s installment of mathematical content will be on the inscribed square problem. This problem is also referred to as the square peg problem, and Toeplitz’ conjecture. This problem though simple, remains unsolved still to today. There is not a million-dollar reward for this problem, but in exchange for the lack of monetary incentive the problem is easy to understand. There have not been a large number of developments in this problem, so the history is brief. In order to compensate a rectangular proof will be mentioned. To format this post, the first part will be the brief history; second will be the problem, and finally the rectangular proof.

It is called the Toeplitz conjecture, because Jewish mathematician, Otto Toeplitz, created it. Otto was born in 1881, and much like his father, and grandfather, he wanted to work in mathematics. He received his doctorate in 1905, and in 1906 started working at Gottingen University. At the time, Gottingen was one of the most prestigious universities in mathematical development. He worked with mathematicians such as Felix Klein, and David Hilbert. During his time there he worked on linear, and quadric functions in high special dimensions. Then in the year 1911 he proposed his conjecture on an inscribed square within a curve.

His conjecture is that on any curve there can be found four points that can construct a square. For example, a circle is probably the easiest to visualize a square inscribed within. In fact, if the square is rotated it can be seen that there can be an infinite number of squares found within a circle. However, this does not only apply to circles, the conjecture says it applies to all Jordan curves. A Jordan curve being a closed curve not having an intersection of any kind. For example, oval shapes, kidney bean shapes, and heart shapes would all be Jordan curves. However, self-intersecting shapes such as a figure eight, or infinity symbol would not be Jordan curves. This also includes all irregular shape variants that do not intersect. This conjecture was purposed a hundred years ago, and still it remains unsolved today. There has not been any progress on the square problem, but it has been proven that this problem can be solved with triangles, and also can be solved with rectangles.

As is the case for most mathematical problems, the solutions tend to be more complicated than the questions. This is the case with the rectangular proof of this problem. Just to be clear, this is not a proof of the inscribed square, but rather an inscribed rectangle. The approach to solving this is to look at a property of rectangles. A rectangle has four points, and two pairs of diagonal lines can be draw within a rectangle. It is important to recognize that the two diagonals have the same length. In fact, if two line segments share a midpoint, and are the same length, a rectangle can be formed from those four corresponding points. So rather than proving that a rectangle can be found, it is easier to prove that two line segments, that share a midpoint, and have the same length, can be found in any Jordan curve. The proof involves mapping the curve in three-dimensional space. Imagine a closed Jordan curve on a horizontal plane, then draw a line connecting two points on the curve. Take the distance between the two points and graph it vertically on the z axis, at the midpoint. Continually doing this over every pair of points on the curve generates a three-dimensional model of the curve. This being hard to visualize, a circle would map to a nice dome shape. At this point it needs to be made clear that the two line segments that are being looked at, need to be different line segments. This is to make sure that a rectangle would be formed, and not just a line. In order to do this the points on the curve must be unordered pairs. This just means that it removes redundancy from our problem. This next part will not be fully explained, but it can be proven. Any closed loop with unordered point pairs maps to a Mobius strip. The outer edge of the Mobius strip, corresponds to the points of the outer edge of the Jordan curve. In order to transform the Mobius strip into the three-dimension curve generated at the midpoints, it requires it to intersect itself. Meaning that there are two pairs of point on the outside edge that correspond in distance and midpoint.


The last part was not easily understandable, with just words. Here is a link to a video that gives a nice visualization to this problem. https://www.youtube.com/watch?v=AmgkSdhK4K8

No comments:

Post a Comment