Sublinear Functional

In linear algebra, a sublinear function (or functional as is more often used in functional analysis), also called a quasi-seminorm or a Banach functional, on a vector space $X$ is a real-valued function with only some of the properties of a seminorm. Unlike seminorms, a sublinear function does not have to be nonnegative-valued and also does not have to be absolutely homogeneous. Seminorms are themselves abstractions of the more well known notion of norms, where a seminorm has all the defining properties of a norm that it is not required to map non-zero vectors to non-zero values.

In functional analysis the name Banach functional is sometimes used, reflecting that they are most commonly used when applying a general formulation of the Hahn–Banach theorem.
The notion of a sublinear function was introduced by Stefan Banach when he proved his version of the Hahn-Banach theorem.

There is also a different notion in computer science, described below, that also goes by the name "sublinear function."

Definitions

Let $X$ be a vector space over a field $\mathbb,$ where $\mathbb$ is either the real numbers $\R$ or complex numbers $\C.$
A real-valued function $p : X \to \R$ on $X$ is called a ' (or a ' if $\mathbb = \R$), and also sometimes called a ' or a ', if it has these two properties:

1. ''Positive homogeneity/Nonnegative homogeneity'': $p(r x) = r p(x)$ for all real $r \geq 0$ and all $x \in X.$
   * This condition holds if and only if $p(r x) \leq r p(x)$ for all positive real $r > 0$ and all $x \in X.$


2. ''Subadditivity/Triangle inequality'': $p(x + y) \leq p(x) + p(y)$ for all $x, y \in X.$
   * This subadditivity condition requires $p$ to be real-valued.

A function $p : X \to \R$ is called ' or ' if $p(x) \geq 0$ for all $x \in X.$
It is a ' if $p(-x) = p(x)$ for all $x \in X.$
Every subadditive symmetric function is necessarily nonnegative.Let $x \in X.$ The triangle inequality and symmetry imply $p(0) = p(x + (- x)) \leq p(x) + p(-x) = p(x) + p(x) = 2 p(x).$ Substituting $0$ for $x$ and then subtracting $p(0)$ from both sides proves that $0 \leq p(0).$ Thus $0 \leq p(0) \leq 2 p(x)$ which implies $0 \leq p(x).$ $\blacksquare$
A sublinear function on a real vector space is symmetric if and only if it is a seminorm.
A sublinear function on a real or complex vector space is a seminorm if and only if it is a balanced function or equivalently, if and only if $p(u x) \leq p(x)$ for every unit length scalar $u$ (satisfying $, u, = 1$) and every $x \in X.$

The set of all sublinear functions on $X,$ denoted by $X^,$ can be partially ordered by declaring $p \leq q$ if and only if $p(x) \leq q(x)$ for all $x \in X.$
A sublinear function is called ' if it is a minimal element of $X^$ under this order.
A sublinear function is minimal if and only if it is a real linear functional.

Examples and sufficient conditions

Every norm, seminorm, and real linear functional is a sublinear function.
The identity function $\R \to \R$ on $X := \R$ is an example of a sublinear function (in fact, it is even a linear functional) that is neither positive nor a seminorm; the same is true of this map's negation $x \mapsto -x.$
More generally, for any real $a \leq b,$ the map
\begin
S_ :\;&& \R &&\;\to \;& \R \\ .3ex && x &&\;\mapsto\;&
\begin
a x & \text x \leq 0 \\
b x & \text x \geq 0 \\
\end \\
\end
is a sublinear function on $X := \R$ and moreover, every sublinear function $p : \R \to \R$ is of this form; specifically, if $a := - p(-1)$ and $b := p(1)$ then $a \leq b$ and $p = S_.$

If $p$ and $q$ are sublinear functions on a real vector space $X$ then so is the map $x \mapsto \max \.$ More generally, if $\mathcal$ is any non-empty collection of sublinear functionals on a real vector space $X$ and if for all $x \in X,$ $q(x) := \sup \,$ then $q$ is a sublinear functional on $X.$

A function $p : X \to \R$ is sublinear if and only if it is subadditive, convex, and satisfies $p(0) \leq 0.$

Properties

Every sublinear function is a convex function: For $t \in$, 1/math>,
:$\begin p(t x + (1 - t) y) &\leq p(t x) + p((1 - t) y) & \text \\ &= t p(x) + (1 - t) p(y) & \text \end$

If $p : X \to \R$ is a sublinear function on a vector space $X$ thenIf $x \in X$ and $r := 0$ then nonnegative homogeneity implies that $p(0) = p(r x) = r p(x) = 0 p(x) = 0.$ Consequently, $0 = p(0) = p(x + (-x)) \leq p(x) + p(-x),$ which is only possible if $0 \leq \max \.$ $\blacksquare$
$p(0) ~=~ 0 ~\leq~ p(x) + p(- x) \qquad \text x \in X,$
which implies that at least one of $p(x)$ and $p(- x)$ must be nonnegative; that is,
$0 ~\leq~ \max \ \qquad \text x \in X.$
Moreover, when $p : X \to \R$ is a sublinear function on a real vector space then the map $q : X \to \R$ defined by $q(x) := \max \$ is a seminorm.

Subadditivity of $p : X \to \R$ guarantees$p(x) = p(y + (x - y)) \leq p(y) + p(x - y),$ which happens if and only if $p(x) - p(y) \leq p(x - y).$ $\blacksquare$
$p(x) - p(y) ~\leq~ p(x - y) \qquad \text x, y \in X$
so if $p$ is also symmetric then the reverse triangle inequality will hold:
$, p(x) - p(y), ~\leq~ p(x - y) \qquad \text x, y \in X.$

Associated seminorm

If $p : X \to \R$ is a real-valued sublinear function on a real vector space $X$ (or if $X$ is complex, then when it is considered as a real vector space) then the map $q(x) := \max \$ defines a seminorm on the real vector space $X$ called the seminorm associated with $p.$
A sublinear function $p$ on a real or complex vector space is a symmetric function if and only if $p = q$ where $q(x) := \max \$ as before.

More generally, if $p : X \to \R$ is a real-valued sublinear function on a (real or complex) vector space $X$ then
$q(x) ~:=~ \sup_ p(u x) ~=~ \sup \$
will define a seminorm on $X$ if this supremum is always a real number (that is, never equal to $\infty$).

Relation to linear functionals

If $p$ is a sublinear function on a real vector space $X$ then the following are equivalent:

1. $p$ is a linear functional.


2. for every $x \in X,$ $p(x) + p(- x) \leq 0.$


3. for every $x \in X,$ $p(x) + p(- x) = 0.$


4. $p$ is a minimal sublinear function.

If $p$ is a sublinear function on a real vector space $X$ then there exists a linear functional $f$ on $X$ such that $f \leq p.$

If $X$ is a real vector space, $f$ is a linear functional on $X,$ and $p$ is a positive sublinear function on $X,$ then $f \leq p$ on $X$ if and only if $f^(1) \cap \ = \varnothing.$

Dominating a linear functional

A real-valued function $f$ defined on a subset of a real or complex vector space $X$ is said to be a sublinear function $p$ if $f(x) \leq p(x)$ for every $x$ that belongs to the domain of $f.$
If $f : X \to \R$ is a real linear functional on $X$ then $f$ is dominated by $p$ (that is, $f \leq p$) if and only if $-p(-x) \leq f(x) \leq p(x) \quad \text x \in X.$
Moreover, if $p$ is a seminorm or some other (which by definition means that $p(-x) = p(x)$ holds for all $x$) then $f \leq p$ if and only if $, f, \leq p.$

Continuity

Suppose $X$ is a topological vector space (TVS) over the real or complex numbers and $p$ is a sublinear function on $X.$
Then the following are equivalent:

1. $p$ is continuous;


2. $p$ is continuous at 0;


3. $p$ is uniformly continuous on $X$;

and if $p$ is positive then we may add to this list:

4. $\$ is open in $X.$

If $X$ is a real TVS, $f$ is a linear functional on $X,$ and $p$ is a continuous sublinear function on $X,$ then $f \leq p$ on $X$ implies that $f$ is continuous.

Relation to Minkowski functions and open convex sets

Relation to open convex sets

Operators

The concept can be extended to operators that are homogeneous and subadditive.
This requires only that the codomain be, say, an ordered vector space to make sense of the conditions.

Computer science definition

In computer science, a function $f : \Z^+ \to \R$ is called sublinear if $\lim_ \frac = 0,$ or $f(n) \in o(n)$ in asymptotic notation (notice the small $o$).
Formally, $f(n) \in o(n)$ if and only if, for any given $c > 0,$ there exists an $N$ such that $f(n) < c n$ for $n \geq N.$
That is, $f$ grows slower than any linear function.
The two meanings should not be confused: while a Banach functional is convex, almost the opposite is true for functions of sublinear growth: every function $f(n) \in o(n)$ can be upper-bounded by a concave function of sublinear growth.

See also

*
*
*
*
*
*
*
*

Notes

References

Bibliography

*
*
*
*
*

{{Topological vector spaces

Articles containing proofs
Functional analysis
Linear algebra
Types of functions