For any $x,y> 0, x\leq y, xy\geq 1,$ and real values of $\theta,$ we can establish easily the inequality $$\left|\frac{1-xe^{i\theta}}{1-ye^{i\theta}}\right|\leq \sqrt{\frac{x}{y}}, $$ for all $y$ and $\theta$ for which $1-ye^{i\theta}\neq 0.$ But my idea is to relax the condition $xy\geq 1.$ Can we have a bound for $\left|\frac{1-xe^{i\theta}}{1-ye^{i\theta}}\right|$ for any $x,y\geq 0, x\leq y.$
2026-04-02 23:28:56.1775172536
Estimating the modulus of a ratio
54 Views Asked by Bumbble Comm https://math.techqa.club/user/bumbble-comm/detail At
1
Due to the symmetry, it suffices to consider $0\leq\theta\leq \pi.$
Denote $O,A,X,Y$ the points corresponding respectively to the numbers $0,\;1,\;xe^{i\theta},\;ye^{i\theta}.$
Assume $0\neq\theta\neq \pi.$ With the use of the law of cosines in $\triangle OAX$ and $\triangle OAY$ we get $$|1-xe^{i\theta}|^2=1+x^2-2x\cos\theta,$$ and $$|1-ye^{i\theta}|^2=1+y^2-2y\cos\theta.$$
If $A,X,Y$ are aligned, the above estimates hold (continuity of $g$ on $\mathbb{R}).$
Note
In the case $xy\geq1$ it holds (easy to verify) $$\left|\frac{1-xe^{i\theta}}{1-ye^{i\theta}}\right|\leq\frac{1+x}{1+y}\leq\sqrt{\frac{x}{y}},$$ thus we have a better upper-estimate.
In notation of arithmetic and geometric means, we have $$\left|\frac{1-xe^{i\theta}}{1-ye^{i\theta}}\right|\leq\frac{AM(x,1)}{AM(y,1)}\leq\frac{GM(x,1)}{GM(y,1)}\quad \text{if}\quad xy\geq 1,$$ and $$\left|\frac{1-xe^{i\theta}}{1-ye^{i\theta}}\right|>\frac{AM(x,1)}{AM(y,1)}>\frac{GM(x,1)}{GM(y,1)}\quad \text{if}\quad xy< 1.$$