W. Rudin, POMA, Ch 1, Ex 6

66 Views Asked by At

Fix $b>1$.

Let's define $B(x)=\{b^t:t\in \mathbb{Q},t\leq x\}$.

Prove that $b^r=\sup B(r)$ when $r$ is rational. Hence it makes sense to define $b^x=\sup B(x)$ for every real $x$.

The question is about last sentence: "Hence it makes sense to define $b^x=\sup B(x)$ for every real $x$."

Can someone explain rigorously how our prove for rational $x$ makes sense define it for real $x$?

I think it is because the real line is complete and that rational numbers are dense in real line. But I dont know how to show it rigorously.

1

There are 1 best solutions below

8
On BEST ANSWER

The fact that "the real line is complete and the rational numbers are dense in the real line" perhaps helps to justify that the definition of $b^x$ over $x \in \mathbb{R}$ matches our visual intuitive picture of what the graph of an exponential function should look like; but that is not what is needed here just for the definition to "make sense" - you are overcomplicating things!


For any $b>0$ and any rational $x \in \mathbb{Q}$, we already have a definition of $b^x$. With this, for any $b>0$ and any real $x \in \mathbb{R}$, we can define a set $B(x)$ by $$ B(x) := \{b^t : t \in \mathbb{Q}, t \leq x\}\text{.} $$ It is easy to check that for any real $x$, the set $B(x)$ is bounded above, and therefore $\,\sup B(x)\,$ is a well-defined real number!

In the case that $x$ is rational, this number $\,\sup B(x)\,$ turns out to be precisely the value of $b^x$. Therefore, it would not create any inconsistency of definitions to introduce a definition of $b^x$ for all real $x$ as $$ b^x := \sup B(x)\text{,} $$ since those $x$-values for which $b^x$ has already been defined (namely, the rationals) give the same value for $b^x$ under our previously-existing definition (defined only for rational $x$) as under our new definition (defined for all real $x$).