The other approach to convergence is filter convergence. Filters are introduced for relational structures
as non empty, upper, and filtered subsets (see [W0]). We need filters on a set
, i.e., filters of
,
the Boolean poset of the set
(see [Y1]).
In other words, filters on a set
are non empty families of subsets of
characterized
by two conditions (see [W7]):
An easy example of a filter on a topological space is the neighborhood system of a given point.
Indeed, a superset of a neighborhood is itself a neighborhood and the intersection of two neighborhoods
is also a neighborhood. The neighborhood system of a point
is convergent to
.
Another example of a filter is the filter of a given net on
. It is a family of all subsets
of
such that
the given net is eventually in
. Closeness on supersets is a direct consequence of the definition of the concept
``eventually in''. For the proof of closeness on the operation of intersection we must
also apply the fact that indices form a directed set. The filter of a net convergent to a point
is also convergent to
.
This is one aspect of equivalence of two convergences.
The opposite conversion - from filters to nets - is more difficult. The net of a given filter is defined in [Y19] as follows:
definition
let S be non empty 1-sorted;
let O be non empty Subset of S;
let F be Filter of BoolePoset O;
func a_net F -> strict non empty NetStr over S means
the carrier of it = {[a, f] where a is Element of S,
f is Element of F: a in f} &
(for i,j being Element of it holds i <= j iff j`2 c= i`2) &
for i being Element of it holds it.i = i`1;
end;
For a proper filter
, the filter of the net of
is equal to
.
Finally, if a proper filter is convergent to a point
, then the net of it is also convergent to
.
This completes the expression of the equivalence of two convergences as it is worked out in [Y19].