# Inequality between taxicab metric and Euclidean metric [closed]

Prove that in $\mathbb R^n$, $d_1 (x,y)\le{\sqrt n}d_2 (x,y)$. Here $d_1(x,y)=\sum|x_i-y_i|$ and $d_2(x,y) = \sqrt{\sum (x_i-y_i)^2}$.

I tried to square both sides and expand but do not think that I can prove that $n\lvert x_i -y_i\rvert<\sum_{i=1}^n \lvert x_i -y_i\rvert$ for any i. Maybe this has something to do with Cauchy-Schwarz but I don't see how to use it.

## closed as unclear what you're asking by DeepSea, user91500, Harish Chandra Rajpoot, Tom-Tom, ClaytonSep 9 '15 at 14:37

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

• Please don't assume that every little notation your book uses is universally recognized by everyone. What exactly do you mean by $d_1, d_2$? – Paul Sinclair Sep 9 '15 at 5:19
• $d_2$ is the euclidean metric – Ricc Sep 9 '15 at 5:20
• Okay, the more obvious one is down. Now what about the other one? – Paul Sinclair Sep 9 '15 at 5:21
• $d_1= \sum_{j=1}^n \lvert x_j - y_j\rvert$ – Ricc Sep 9 '15 at 5:22

hint: $d_1 \leq \sqrt{n}\cdot d_2$ by CS inequality precisely.
• wow I suddenly see. So basically we let $b_n$ in CS inequality be 1, is that right? – Ricc Sep 9 '15 at 5:27
Consider $f,g$ defined on the set $\{(x_1,x_2,\ldots, x_n)\in \mathbb{R}^n : x_i\geq 0 \quad\forall i\}$ by $$f(x_1,x_2,\ldots, x_n)= \sum_{i=1}^n x_i, \text{ and } g(x_1,x_2,\ldots, x_n) = \sum_{i=1}^n x_i^2$$ and we maximize $f$ under the constraint that $g = 1$. Lagrange multipliers will tell you that this is maximum when $$x_i = \frac{1}{\sqrt{n}} \quad\forall 1\leq i\leq n \Rightarrow f(x_1,x_2,\ldots, x_n) = \sqrt{n}$$ Hence, for any $x\in \mathbb{R}^n$, consider $$y = \frac{x}{d_2(x,y)}$$ Then $d_2(y,0) = 1$, whence $d_1(y,0) \leq \sqrt{n}$ and so $$d_1(x,0) \leq \sqrt{n}d_2(x,0)$$ Now use the fact that $d_j(x,y) = d_j(x-y,0)$ for both $j=1, 2$.