Fix a metrizable space X and a metric d on X that gives the topology. Fix also a closed set S in X. For x in X, define
d(x,S) = inf {d(x,s): s in S}
and define for n = 1, 2, 3, ... the subsets
S_n = {x in X: d(x,S) < 1/n}
of X. I claim (1) for all n the set S_n is open, and (2) S is the intersection of all of the sets S_n. These two facts together show that S is a G-delta set.
Fix n and let's show that S_n is open. Fix y in S_n. By definition of S_n we have d(y,S) < 1/n, so there is s in S with d(y,s) < 1/n. The number e = 1/n - d(y,s) is positive. I claim the open ball in X with center y and radius e is entirely contained in S_n. But this is clear: if x in X satisfies d(x,y) < e then
and hence d(x,S) < 1/n, and hence x in S_n, as desired. This shows that S_n is open.
It is clear that if t is in S then d(t,S) = 0 (because the set {d(t,s): s in S} includes the number d(t,t) = 0) and hence that t is in each of the sets S_n. This shows that S is contained in the intersection of the sets S_n. Conversely suppose that x in X is in all of the sets S_n. Construct a sequence y_n in S as follows: fixing n, from the fact that x is in S_n we know d(x,S) < 1/n, and hence that there is some element y of S with d(x,y) < 1/n; choose one and call it y_n. The sequence thus defined satisfies d(x,y_n) < 1/n for all n. It follows that the sequence (y_n) converges to x (since if e > 0 is given, choosing N so that 1/N < e, one has that d(x,y_n) < 1/n < 1/N < e for all n >= N) and hence that x is in the closure of the set S, and since S is closed, that x is in S. This concludes the proof.
The converse is not true although I had to go to the web for examples. Wikipedia tells me that spaces where every closed set is G-delta are called "G-delta spaces". It lists various properties of these spaces and provides an example of a space, the Sorgenfrey line, which is a G-delta space but not metrizable.
each place it particularly is administered via the government, federal, state, county and native is closed. you're able to circulate to the submit place of work in case you have something to deliver out. because of the fact maximum submit workplaces have those blue mailboxes exterior. happy Easter to you and yours. :-)
Answers & Comments
Verified answer
Fix a metrizable space X and a metric d on X that gives the topology. Fix also a closed set S in X. For x in X, define
d(x,S) = inf {d(x,s): s in S}
and define for n = 1, 2, 3, ... the subsets
S_n = {x in X: d(x,S) < 1/n}
of X. I claim (1) for all n the set S_n is open, and (2) S is the intersection of all of the sets S_n. These two facts together show that S is a G-delta set.
Fix n and let's show that S_n is open. Fix y in S_n. By definition of S_n we have d(y,S) < 1/n, so there is s in S with d(y,s) < 1/n. The number e = 1/n - d(y,s) is positive. I claim the open ball in X with center y and radius e is entirely contained in S_n. But this is clear: if x in X satisfies d(x,y) < e then
d(x,s) <= d(x,y) + d(y,s) < e + d(y,s) = 1/n - d(y,s) + d(y,s) = 1/n
and hence d(x,S) < 1/n, and hence x in S_n, as desired. This shows that S_n is open.
It is clear that if t is in S then d(t,S) = 0 (because the set {d(t,s): s in S} includes the number d(t,t) = 0) and hence that t is in each of the sets S_n. This shows that S is contained in the intersection of the sets S_n. Conversely suppose that x in X is in all of the sets S_n. Construct a sequence y_n in S as follows: fixing n, from the fact that x is in S_n we know d(x,S) < 1/n, and hence that there is some element y of S with d(x,y) < 1/n; choose one and call it y_n. The sequence thus defined satisfies d(x,y_n) < 1/n for all n. It follows that the sequence (y_n) converges to x (since if e > 0 is given, choosing N so that 1/N < e, one has that d(x,y_n) < 1/n < 1/N < e for all n >= N) and hence that x is in the closure of the set S, and since S is closed, that x is in S. This concludes the proof.
The converse is not true although I had to go to the web for examples. Wikipedia tells me that spaces where every closed set is G-delta are called "G-delta spaces". It lists various properties of these spaces and provides an example of a space, the Sorgenfrey line, which is a G-delta space but not metrizable.
each place it particularly is administered via the government, federal, state, county and native is closed. you're able to circulate to the submit place of work in case you have something to deliver out. because of the fact maximum submit workplaces have those blue mailboxes exterior. happy Easter to you and yours. :-)