Ma · Mathematics

What Is the Difference Between Supremum and Maximum?

Published By the 9to5Science editorial team

At a glance

  1. 1. Check upper bounds

    An upper bound is any number at least as large as every element of the set.

  2. 2. Find the least bound

    The supremum is the smallest upper bound; it may or may not belong to the set.

  3. 3. Check membership

    A maximum must belong to the set and be at least as large as every other element.

  4. 4. Compare the two

    If a nonempty set has a maximum, that element is also its supremum.

The difference is about whether the bound must be an element of the set. A supremum is the least number that is at least as large as every element; a maximum is an element of the set that is at least as large as every other element.

Thus, a maximum is always a supremum, but a supremum need not be a maximum. The distinction matters especially when a set approaches a boundary that it does not include.

Definitions

Let AA be a nonempty subset of the real numbers that is bounded above. A number uu is an upper bound for AA if x≤ux \leq u for every x∈Ax \in A. The supremum, written sup⁡A\sup A, is the least of all upper bounds. Equivalently, it satisfies two conditions:

  1. x≤sup⁡Ax \leq \sup A for every x∈Ax \in A.
  2. If vv is any upper bound for AA, then sup⁡A≤v\sup A \leq v.

A number mm is the maximum of AA, written max⁡A\max A, if m∈Am \in A and x≤mx \leq m for every x∈Ax \in A. The membership requirement is the key difference.

In R\mathbb{R}, every nonempty set that is bounded above has a supremum, by the completeness property of the real numbers. It does not necessarily have a maximum.

When the maximum exists

If m=max⁡Am = \max A, then mm is an upper bound for AA. It is also the least upper bound: any upper bound must be at least as large as the element mm itself. Therefore,

max⁡A=sup⁡A\max A = \sup A

whenever AA has a maximum.

The converse does not always hold. A supremum can be a boundary value that the set gets arbitrarily close to but never reaches. If the supremum belongs to the set, however, it is the maximum: sup⁡A∈A\sup A \in A implies max⁡A=sup⁡A\max A = \sup A.

Worked examples

An open interval

Take A=(0,1)A = (0,1), the set of real numbers strictly between 0 and 1. Every element of AA is less than 1, so 1 is an upper bound. No smaller number is an upper bound: if u<1u < 1, there is an element of (0,1)(0,1) greater than uu. Thus,

sup⁡A=1.\sup A = 1.

But 1∉A1 \notin A, so there is no maximum. In particular, 0.999 is not a maximum: there are elements of the interval larger than 0.999.

A closed interval

For B=[0,1]B = [0,1], the endpoint 1 belongs to the set and is at least as large as every element. Therefore,

max⁡B=sup⁡B=1.\max B = \sup B = 1.

The only change from the previous example is that the endpoint is included.

A sequence of reciprocals

Let C={1/n:n∈N, n≥1}C = \{1/n : n \in \mathbb{N},\ n \geq 1\}. Its elements begin 1,1/2,1/3,1/4,…1, 1/2, 1/3, 1/4, \ldots. The largest element is 1, which belongs to the set. Hence,

max⁡C=sup⁡C=1.\max C = \sup C = 1.

By contrast, if the set starts at n=2n=2, then D={1/n:n∈N, n≥2}D = \{1/n : n \in \mathbb{N},\ n \geq 2\} has maximum and supremum both equal to 1/21/2. The sequences tend to 0, but 0 is neither the supremum nor the maximum of either set; their maxima are the first terms, 1 and 1/21/2.

A practical test for a maximum

For a nonempty, bounded-above set AA, first find or identify s=sup⁡As = \sup A. Then check whether s∈As \in A:

  • If s∈As \in A, then s=max⁡As = \max A.
  • If s∉As \notin A, then AA has no maximum.

For example, for A=(0,1)A = (0,1), the supremum is 1 but 1 is excluded, so no maximum exists. For B=[0,1]B = [0,1], the supremum is again 1 and is included, so it is the maximum.

Common mistakes

  • Treating any upper bound as the supremum. For (0,1)(0,1), 2 is an upper bound, but it is not the least upper bound; the supremum is 1.
  • Assuming a supremum must be in the set. The definition requires it to be the least upper bound, not a member. The set (0,1)(0,1) is a standard counterexample.
  • Calling an excluded endpoint a maximum. A maximum must be an element of the set. If the interval is open at its upper endpoint, that endpoint cannot be the maximum.
  • Assuming every set has a finite supremum or a maximum. A set that is unbounded above, such as the positive integers, has no finite real supremum and no maximum. The empty set also has no maximum; its supremum depends on the convention being used, so the definitions above assume a nonempty set.

When you compare the terms, ask two questions: is the number an upper bound, and does it belong to the set? The least upper bound answers the first question and the maximum must answer both.

References

  1. Walter Rudin, Principles of Mathematical Analysis

Spotted an error? Send a correction. Read how we write and check articles in our editorial policy.