scientific article; zbMATH DE number 2165651

The celebrated Erdős-Mordell inequality states that if \(P\) is a point inside triangle \(ABC\) at distances \(x_1\), \(x_2\), and \(x_3\) from the vertices \(A\), \(B\), and \(C\), and distances \(d_1\), \(d_2\), and \(d_3\) from the opposite sides, then \(x_1 + x_2 + x_3 \geq 2(d_1 + d_2 + d_3)\). This inequality has first appeared as a problem proposal by \textit{Paul Erdős} in [Problem 3740, Am. Math. Mon. 42, 396 (1935)] and was solved by \textit{L. J. Mordell} and others in [Am. Math. Mon. 44, 252-254 (1937)]. Since then, the inequality has attracted much attention and several different proofs. An attempt to extend this inequality to points outside \(ABC\) (with the obvious understanding that the \(d_i\) are signed distances, i.e., trilinear coordinates) is made by \textit{H. Demir} in [Problem E2462, Am. Math. Mon. 81, 281 (1974)], a problem that was announced by \textit{C. W. Dodge} in [Am. Math. Mon. 92, 360 (1985)] to have attracted no solutions. Later, \textit{C. W. Dodge} published his own solution in [Crux Mathematicorum 10, 274-281 (1984)]. In this paper, the author proves the stronger inequality \(x_1 + x_2 + x_3 \geq u_1 d_1 + u_2 d_2 + u_3 d_3\), where \(u_1 = \frac{b}{c} + \frac{c}{b}\), and where \(u_2\) and \(u_3\) are the iterates under permuting. Here, \(a\), \(b\), and \(c\) are the side-lengths of triangle \(ABC\). He also proves that this stronger inequality holds for points ouside \(ABC\) as well. In both forms of the inequality, equality holds only when \(ABC\) is equilateral and when \(P\) is its center. The author is obviously unaware of the discussion on pages 86-87 of \textit{N. D. Kazarinoff}'s [Geometric Inequalities (New Mathematical Library 4, Math. Assoc. America, Washington, D. C.) (1961)], where this strong form is proved. It is also worth mentioning that the afore-mentioned proof by Mordell does also include this stronger form. The short, simple, and non-trigonometric proof under review, however, will make this strong form of the inequality better known, as it deserves to be. It would also be pleasant if the rational functions \(u_1= \frac{b}{c}+\frac{c}{b}\), etc., turn out to be, in some sense, the best possible rational functions in \(a\), \(b\), and \(c\) for which the inequality \(\sum x_i \geq \sum u_id_i\) holds.

Recommendations

Cited in

(13)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4675586)