I am reading Rudin's real and complex analysis. On page 14 we see $\underset{n} {\text{sup }} f_n$, and on page 15 we see max$\{f,g\}$. These two are obviously different. I searched some answers and found an example:

sup $\{f,g\} = f(x)$ wehn $f(x)\geq g(x)$ and $g(x)$ when $f(x) < g(x)$

But then I saw max$\{f,g\}$ which makes me confused. Are they the same? What are the differences?

Can someone also give some explanation of $(\underset{n\to \infty}{\text{lim sup }} f_n)(x)$? How to visualize/understand it?

  • 2
    $\begingroup$ Not to be too condescending here but I need to make sure you know that Real and Complex Analysis is a graduate textbook and Rudin's undergraduate book is Principles of Mathematical Analysis. I'm pointing this out because your question is the sort of question that should have been addressed in a prerequisite course for this textbook. $\endgroup$ – Trevor Gunn Feb 11 '18 at 5:05

Supremum and Maximum

Let $A$ be a set of real numbers.

If $x$ is a real number such that $a \leq x$ for all $a \in A$, then $x$ is called an upper bound for $A$. Example: $1$ is an upper bound for the interval $(0,1)$. So is $2$.

If $x$ is an upper bound for $A$ and $x$ belongs to $A$, then $x$ is called the maximum of $A$ and we write $x = \max A$. In other words, the maximum of $A$ is the element of $A$ that is larger than every other element of $A$. Not every set has a maximum. Examples: The interval $(0,1)$ has no maximum; The finite set $\{1,2,3,4\}$ has $4$ as its maximum. If the maximum of $A$ exists, then it is unique. This justifies the use of "the maximum" rather than "a maximum."

If $x$ is an upper bound for $A$ and $x$ is smaller than every other maximum for $A$, then $x$ is called the supremum of $A$ (or the least upper bound for $A$) and we write $x = \sup A$. In symbols, $\sup A$ is the number such that (1) $a \leq \sup A$ for all $a \in A$ and (2) if $x'$ is a real number such that $a \leq x'$ for all $a \in A$, then $\sup A \leq x'$.

The supremum of $A$ is unique if it exists. The Dedekind Completeness Property of the Real Numbers is: If $A$ has at least one upper bound, then the supremum of $A$ exists.

If $\max A$ exists, then $\sup A$ exists and $\max A = \sup A$. Example: $A=(0,1]$ has $\max A = \sup A = 1$.

The supremum may exist even when the maximum does not. Example: $A=(0,1)$ has no maximum, but $\sup A = 1$.

Sometimes neither the maximum nor the supremum exist. Example: The interval $(0,+\infty)$ has no maximum and no supremum.

If $\sup A$ exists and belongs to $A$, then $\max A = \sup A$. Example: $A=(0,1]$ has $\max A = \sup A = 1$.

To put things another way, a real number is called...

(i) an upper bound for $A$ if it is greater than or equal to every element of $A$.

(ii) the supremum of $A$ if it is an upper bound for $A$ and it is less than or equal to every upper bound of $A$.

(iii) the maximum if $A$ if it is the supremum of $A$ and it belongs to $A$.

Limit Superior ($\limsup$)

If $(a_n)$ is a sequence of real numbers that is bounded, we can define a new sequence $b_n = \sup \{ a_k : k \geq n\}$. Note that $(b_n)$ is a decreasing sequence. We define $$\limsup a_n = \lim_{n \to \infty} b_n = \lim_{n \to \infty} \sup \{ a_k : k \geq n\}.$$ A useful fact to help understand the meaning of $\limsup a_n$ is that
$$ \limsup a_n = \sup S $$ where $S$ is the set of all real numbers $x$ such that $\lim_{k \to \infty} a_{n_k} = x$ for some convergent subsequence $(a_{n_k})$ of the sequence $(a_n)$. In fact, $\sup S$ belongs to $S$, so $$ \limsup a_n = \max S $$ Thus, you can think of $\limsup a_n$ as the largest limit for all convergent subsequences of $(a_n)$.

Example: The sequence $(1,\frac{1}{2},1,\frac{1}{3},1,\frac{1}{4},1,\ldots)$ does not converge. It has a subsequence $(1,\frac{1}{2},\frac{1}{3},\frac{1}{4},\ldots)$ that converges to $0$. It has another subsequence $(1,1,1,1,\ldots)$ that converges to $1$. It has infinitely many other convergent subsequences, but all of these either converge to $0$ or converge to $1$. Therefore $S = \{0,1\}$. Hence $\limsup a_n = \sup S = \max S = 1$.

For a sequence of real numbers that is bounded, the $\limsup$ always exists as a finite real number, even if the limit of the sequence does not.

If $(a_n)$ is a sequence of real numbers that is not bounded above, then we define $\limsup a_n = +\infty$. If $(a_n)$ is a sequence of real numbers such that $\lim_n a_n = -\infty$, then we define $\limsup a_n = -\infty$.

Given a sequence of functions $f_n:E \to \mathbb{R}$, for each fixed $x \in E$, the sequence $(f_n(x))$ is a sequence of real numbers, so we can talk about $\limsup f_n(x)$. We define the function $\limsup f_n: E \to [-\infty,\infty]$ by setting $$ (\limsup f_n)(x) = \limsup f_n(x). $$ for each fixed $x \in E$.

Note that since we have not specified that the sequence of functions $f_n$ is bounded, it could be that $\limsup f_n(x) = \pm \infty$. That is why the range of the function $\limsup f_n$ is given as $[-\infty,\infty]$.

  • $\begingroup$ $b_{n+1} \leq b_n$ because $\{a_k : k \geq n+1\} \subseteq \{a_k : k \geq n\}$ $\endgroup$ – LucasSilva Feb 17 '18 at 5:43

"max" stands for "maximum" the greatest value in a set. "sup" stands for "supremum" or "least upper bound" of the values in a set. When the set is finite, these mean the same thing, and which one you use is a matter of state. When the set is infinite, it may not have a maximum. For example, what is $\max \{ x | 0 < x < 1\}?$ Obviously, there is no maximum, but the supremum is $1$. So for infinite sets, you shouldn't use "max" unless somehow you know that the set has a maximum. For example, we know that a continuous function attains its maximum on a closed interval, so it would be okay to use "max" in that case, though "sup" would be just as good.


If $x = \max A$ for some set $A \subset \mathbb R$, then $x \ge y$ for any $y \in A$, and $x \in A$.

But the $\sup A$ does not need to be in $A$. By definition if $x = \sup A$, for some set $A \subset \mathbb R$, then $ x \ge y$ for any $y \in A$, but $x \le y$ if $y$ is an upper bound of $A$.


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.