Applied Mathematical Sciences, Vol. 10, 2016, no. 3, 127-135 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/10.12988/ams.2016.511704 Geometric Construction of an llipse from Its Moments Brian J. M c Cartin Applied Mathematics, Kettering University 1700 University Avenue, Flint, MI 48504-6214 USA Copyright c 2015 Brian J. M c Cartin. This article is distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract This paper offers a trio of uclidean constructions of an ellipse given its first moments and second central moments. The crux of these constructions lies in the graphical construction of Carlyle [4, p. 99] and Lill [2] for the roots of a quadratic and the Pappus-uler problem [12] of the construction of the principal axes of an ellipse from a pair of its conjugate diameters. Mathematics Subject Classification: 51M15, 51N20 Keywords: geometric construction, conic section, ellipse 1 Preliminaries An ellipse is uniquely determined by its first moments M x := 1 x dxdy, M y := 1 A A together with its second central moments M xx := 1 A M xy := 1 A y dxdy, (1) (x M x ) 2 dxdy, (2) (x M x )(y M y ) dxdy, (3)
128 B. J. McCartin M yy := 1 A (y M y ) 2 dxdy, (4) where A is the area of the ellipse. Of course, the inequalities M xx > 0, M yy > 0 and M 2 xy < M xx M xx must be satisfied. Specifically [8], the equation of the required ellipse is given by the positive definite quadratic form [ ] [ M x Mx y M xx M xy y M xy M yy and the elliptical area may be expressed as ] 1 [ x Mx y M y ] = 4, (5) A = 4π M xx M yy M 2 xy. (6) The slopes of the principal axes of the ellipse are the roots of the quadratic m 2 + M xx M yy M xy m 1 = 0, (7) so that m major = (M yy M xx ) + (M yy M xx ) 2 + 4M 2 xy 2M xy, (8) m minor = (M yy M xx ) (M yy M xx ) 2 + 4M 2 xy 2M xy, (9) since m major has the same sign as M xy [6]. The corresponding squared magnitudes of the principal semiaxes [15, p. 158] are the roots of the quadratic σ 2 4(M xx + M yy ) σ + 16(M xx M yy M 2 xy) = 0, (10) so that σ 2 major = 2 [ (M xx + M yy ) + (M xx M yy ) 2 + 4M 2 xy], (11) σ 2 minor = 2 [ (M xx + M yy ) (M xx M yy ) 2 + 4M 2 xy]. (12) quation (5) may be expanded thereby yielding M yy (x M x ) 2 2M xy (x M x )(y M y ) + M xx (y M y ) 2 = 4(M xx M yy M 2 xy) (13) as the equation of the ellipse. Setting y = M y yields the corresponding horizontal points of intersection ( Mx ± 2 M xx M 2 xy/m yy, M y ), (14)
Construction of an ellipse from moments 129 while setting x = M x yields the corresponding vertical points of intersection ( Mx, M y ± 2 M yy M 2 xy/m xx ). (15) The points of vertical and horizontal tangency are [7], respectively, ( M x ± 2 M xx, M y ± 2 M ) xy ; M x ± 2 M xy, M y ± 2 M yy. (16) Mxx Myy A 2 A 1 Figure 1: Conjugate Semidiameters With reference to Figure 1, recall that two semidiameters are conjugate to one another if they are parallel to each others tangents [8]. For this to occur, it is necessary and sufficient that their slopes m and s satisfy the symmetric conjugacy condition: [7] M xx (m s) M xy (m + s) + M yy = 0. (17) Thus, the semidiameter to a point of vertical/horizontal tangency is conjugate to the semidiameter to a point of vertical/horizontal intersection, respectively. (See Figures 4/5, Left.) In the succeeding sections, a trio of straightedge-compass constructions of an ellipse,, given its moments, M x, M y, M xx, M xy, M yy, will be developed. Strictly speaking, such a uclidean construction is not possible for the ellipse.
130 B. J. McCartin However, given the principal axes of the ellipse, several methods are available for drawing any number of points on the ellipse [3]. Thus, the phrase construct the ellipse will be interpreted as construct the principal axes of the ellipse. Certain primitive arithmetic/algebraic uclidean constructions [1, pp. 121-122], as next described, will be required. Proposition 1 (Primitive Constructions) Given two line segments of nonzero lengths a and b, one may construct line segments of length a + b, a b, a b, a/b and a using only straightedge and compass. 2 The Carlyle-Lill Construction m major (0,1) m minor (M yy M xx )/M xymmajor m minor (M x,m y ) 1 ((M yy M xx )/M xy, 1) Figure 2: Determination of Axes Directions Because quations (8-9) involve only the operators +,,, /,, application of Proposition 1 would provide an immediate construction of the directions of the principal axes. Furthermore, since quations (11-12) also involve only this same set of operators, an additional application would yield the squared magnitudes of the principal semiaxes thereby leading to their complete construction. Clearly, an approach less reminiscent of a sledgehammer is to be preferred. Instead, the following beautiful construction due to Carlyle [4, p. 99] and Lill [2, pp. 355-356] will be invoked. Proposition 2 (Graphical Solution of a Quadratic quation) Draw a circle having as diameter the line segment joining (0, 1) and (a, b). The abscissae of the points of intersection of this circle with the x-axis are the roots of the quadratic x 2 ax + b = 0.
Construction of an ellipse from moments 131 2 (4(M xx +M yy ),16(M xx M yy M xy )) σ major σ minor (0,1) + 2 + 2 σ minor σ major Figure 3: Determination of Semiaxes Magnitudes A double application of Proposition 2, in concert with quations (7) and (10), produces the following far more elegant construction. Construction 1 (Double Carlyle-Lill Construction) Given the first and second moments M x, M y, M xx, M xy, M yy : 1. According to quation (7), apply the Carlyle-Lill construction, Proposition 2, with (a, b) = ( Myy Mxx M xy, 1) to construct the directions of the principal axes. (See Figure 2.) 2. According to quation (10), apply the Carlyle-Lill construction, Proposition 2, with (a, b) = (4(M xx + M yy ), 16(M xx M yy M 2 xy)) to construct the squared magnitudes of the principal semiaxes. (See Figure 3, Left.) 3. Apply the -operator, Proposition 1, to these squared magnitudes in order to construct the magnitudes, {σ major, σ minor }, of the principal semiaxes emanating from the centroid, (M x, M y ). (See Figure 3, Right.) 4. From the principal semiaxes, { σ major, σ minor }, a variety of methods [3] may now be employed to construct any number of points lying on the ellipse : r(t) = cos θ σ major + sin θ σ minor (0 θ < 2π).
132 B. J. McCartin Figure 4: Vertical Tangent Construction 3 Pappus-uler Problem The following problem was first formulated and solved by Pappus in the 4 th Century, although uler gave the first proof of its validity only in the 18 th Century [9]. Proposition 3 (Pappus-uler Problem) Given two conjugate diameters of an ellipse, to find the principal axes in both position and magnitude. As uler s demonstration of Pappus construction was purely synthetic in nature, McCartin subsequently provided both a purely vector analytic demonstration [10] as well as an independent matrix-vector analytic demonstration [11]. Since that time, many alternative solutions have been devised [5, 14]. Most recently, McCartin has provided a matrix-vector analytic construction [12] which will next be utilized, in tandem with some prior observations, to provide a pair of related constructions of an ellipse from its given moments. As previously noted, the semidiameter of a point of vertical intersection is conjugate to the semidiameter to a point of vertical tangency. Thus, by quations (15) and (16), the vectors 0, 2 M yy Mxy/M 2 xx ; 2 M xx, 2 M xy (18) Mxx define a pair of conjugate semidiameters emanating from the centroid, (M x, M y ), which may be utilized in the Pappus-uler problem, Proposition 3, to construct the principal axes of the ellipse. (See Figure 4.)
Construction of an ellipse from moments 133 Figure 5: Horizontal Tangent Construction Construction 2 (Vertical Tangent Construction) Given the first and second moments M x, M y, M xx, M xy, M yy, apply the McCartin algorithm [12] to the pair of conjugate semidiameters 0, 2 M yy Mxy/M 2 xx ; 2 M xx, 2 M xy Mxx thereby constructing the principal axes of the ellipse in direction and magnitude emanating from the centroid, (M x, M y ). Likewise, the semidiameter of a point of horizontal intersection is conjugate to the semidiameter to a point of horizontal tangency. Thus, by quations (14) and (16), the vectors 2 M yy Mxy/M 2 xx, 0 ; 2 M xy, 2 M yy (19) Myy define a pair of conjugate semidiameters emanating from the centroid, (M x, M y ), which may be utilized in the Pappus-uler problem, Proposition 3, to construct the principal axes of the ellipse. (See Figure 5.) Construction 3 (Horizontal Tangent Construction) Given the first and second moments M x, M y, M xx, M xy, M yy, apply the McCartin algorithm [12] to the pair of conjugate semidiameters 2 M yy Mxy/M 2 xx, 0 ; 2 M xy, 2 M yy Myy
134 B. J. McCartin thereby constructing the principal axes of the ellipse in direction and magnitude emanating from the centroid, (M x, M y ). 4 Postscript In the foregoing, a trio of uclidean constructions for an ellipse given its first moments and second central moments was developed. The first such construction relied heavily upon the Carlyle-Lill graphical construction of the roots of a quadratic equation. The remaining constructions hinged upon the many available solutions to the Pappus-uler problem for the construction of the principal axes of an ellipse given any pair of its conjugate diameters [5, 9, 10, 11, 12, 14]. While the present paper was focused on continuous distributions of mass, given the appropriate first and second moments, the very same constructions are applicable to the construction of the concentration ellipse [6] associated with a discrete distribution of points. Furthermore, this observation permits the unified construction of various lines of regression [13]. References [1] R. Courant and H. Robbins, Geometrical Constructions, Chapter III in What Is Mathematics?, 117-164, Oxford University Press, New York, 1941. [2] L.. Dickson, Constructions with Ruler and Compasses; Regular Polygons in Monographs on Topics of Modern Mathematics Relevant to the lementary Field (J. W. A. Young, ditor), 353-386, Dover, New York, 1955. [3] J. W. Downs, Practical Conic Sections: The Geometric Properties of llipses, Parabolas and Hyperbolas, Dover, New York, 2003 (1993). [4] H. ves, An Introduction to the History of Mathematics, Sixth dition, HBJ-Saunders, New York, 1990. [5] J.. Hofmann and H. Wieleitner, Zur Geschichte der sog. Rytzschen Achsenkonstruktion einer llipse aus einem Paar konjugierter Durchmesser, Nieuw Archief voor Wiskunde (Amsterdam), Series 2, Part 16, 1930, 5-22. [6] B. J. McCartin, A Geometric Characterization of Linear Regression, Statistics, 37 (2003), no. 2, 101-117. http://dx.doi.org/10.1080/0223188031000112881
Construction of an ellipse from moments 135 [7] B. J. McCartin, Oblique Linear Least Squares Approximation, Applied Mathematical Sciences, 4 (2010), no. 58, 2891-2904. [8] B. J. McCartin, A Matrix Analytic Approach to the Conjugate Diameters of an llipse, Applied Mathematical Sciences, 7 (2013), no. 36, 1797-1810. [9] B. J. McCartin, On uler s Synthetic Demonstration of Pappus Construction of an llipse from a Pair of Conjugate Diameters, International Mathematical Forum, 8 (2013), no. 22, 1049-1056. http://dx.doi.org/10.12988/imf.2013.3465 [10] B. J. McCartin, A Vector Analytic Demonstration of Pappus Construction of an llipse from a Pair of Conjugate Diameters, Applied Mathematical Sciences, 8 (2014), no. 20, 963-973. http://dx.doi.org/10.12988/ams.2014.312725 [11] B. J. McCartin, A Matrix-Vector Analytic Demonstration of Pappus Construction of an llipse from a Pair of Conjugate Diameters, Applied Mathematical Sciences, 9 (2015), no. 14, 679-687. http://dx.doi.org/10.12988/ams.2015.4121035 [12] B. J. McCartin, A Matrix-Vector Analytic Construction for the Pappus- uler Problem, Applied Mathematical Sciences, 10 (2016), no. 1, 1-11. http://dx.doi.org/10.12988/ams.2016.510654 [13] B. J. McCartin, Geometric Constructions Relating Various Lines of Regression, Applied Mathematical Sciences, 10 (2016), to appear. [14] Z. Nádeník, O konstrukcích os elipsy z jejích sdru zených pr umĕr u (metodicko-historický p ríspĕvek), in M. Ka sparová and Z. Nádeník (ditors), Jan Sobotka (1862-1931) (Czech), Matfyzpress, Praha, 2010, 127-173. [15] G. Salmon, A Treatise on Conic Sections, Sixth dition, Chelsea, New York, 1954 (1879). Received: December 1, 2015; Published: January 14, 2016