I am still unsure as to how to go about proving if something is a metric space or if a specified distance function defines a metric space. I am attempting to tackle the following and would like any tips\corrections if possible. I know that a distance function must satisfy the following:

  1. $d(x,y)\geq 0$ (equals $0$ if $x=y$)
  2. $d(x,y)=d(y,x)$
  3. $d(x,z) \leq d(x,y) + d(y,z)$

Suppose $d: X \times X \to \mathbb{R}$ is a distance function. Are the following also distance function $\rho: X \times X \to R$:

  1. $\rho(x,y)= (d(x,y))^2$
  2. $\rho(x,y)= (d(x,y))^{1/2}$
  3. $\rho(x,y)= 3d(x,y)$
  4. $\rho(x,y)= (d(x,y))^{1/2} + 2d(x,y)$


  1. Clearly the first two properties hold. The triangle inequality comes down to showing if $d(x,z)^2 \leq d(x,y)^2 + d(y,z)^2$. We run into a contradiction if we let $X$ be the real line and consider the points $x=1, z=-1, y=0$.
  2. Not quite sure
  3. Clearly, the first two properties hold. When we face the triangle inequality, we can factor out the $3$ and our inequality is our usual triangle inequality which we assume holds since we assume $d$ is a distance function.
  4. Because of $2$ I am unsure about it. I believe it will\will not hold depending on this problem.

For 1., let $X$ be the real line with the usual metric, and take $x=-1$, $z=1$, and $y=0$.

For 2., $$\rho(x,z) \le \sqrt{d(x,y)+d(y,z)} \le \sqrt{d(x,y)} + \sqrt{d(y,z)} = \rho(x,y) + \rho(y,z).$$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.