A topological vector space is a vector space over
or
that is also a Hausdorff space and for which vector
addition and scalar multiplication are
jointly continuous. Every normed
space, including each Banach space and Hilbert
space, is a topological vector space in its norm
topology.
The choice of topology determines the meaning of convergence of functions.
For example, on a measure space , convergence in
the norm of
means
|
(1)
|
It implies convergence of the integrals, but the converse need not hold. The equivalence
classes of Lebesgue integrable functions
form the Banach space , the
case of the Lp-spaces.
For pointwise convergence, let be a nonempty set and let
or
. On the vector space
of all functions
,
each
defines a seminorm
|
(2)
|
These seminorms generate the product topology, in which iff
for every
(Tao, p. 119). If
is finite, this topology
is induced by the supremum norm
. If
is infinite, no norm
induces this topology on the full function
space
. A basic neighborhood
of zero restricts only finitely many coordinates and contains
for every
,
for some nonzero function
supported at an unrestricted point. Such a neighborhood
cannot be contained in a norm unit
ball.
For a topological space , the topology of uniform
convergence on compact sets on
, the vector space
of scalar-valued continuous functions, is
generated by the seminorms
|
(3)
|
where
ranges over the nonempty compact
subsets of
(Tao, pp. 119-120).