A FuriPera theorem in Hausdorff topological spaces for acyclic maps
International Journal of
A FURIPERA THEOREM IN HAUSDORFF TOPOLOGICAL SPACES FOR ACYCLIC MAPS
RAVI P. AGARWAL 0
JEWGENI H. DSHALALOW 0
DONAL O'REGAN 0
= W : Fix F ? ? ?F ? (W 0
0 Donal O'Regan: Department of Mathematics, National University of Ireland , Galway , Ireland Email address:
We present new FuriPera theorems for acyclic maps between topological spaces. 1. Introduction. In this paper, we present new FuriPera theorems [6, 7] for acyclic maps between Hausdorff topological spaces. The main result in our paper is based on a new LeraySchauder alternative [1] for such maps which in turn is based on the notion of compactly nullhomotopic. We first recall some results and ideas from the literature. Let X and Z be subsets of Hausdorff topological spaces. We will consider maps F : X ? K(Z); here K(Z) denotes the family of nonempty compact subsets of Z. A nonempty topological space is said to be acyclic if all its reduced C?ech homology groups over the rationals are trivial. Now F : X ? K(Z) is acyclic if F is upper semicontinuous with acyclic values. Suppose X and Z are topological spaces. Given a class of maps, (X, Z) denotes the set of maps F : X ? 2Z (nonempty subsets of Z) belonging to , and c the set of finite compositions of maps in . We let

where Fix F denotes the set of fixed points of F.
The class of maps is defined by the following properties:
(i) contains the class of singlevalued continuous functions;
(ii) each F ? c is upper semicontinuous and compact valued;
(iii) Bn ? ( c) for all n ? {1, 2, . . . }; here Bn = {x ? Rn : x ? 1}.
Next we consider the class c?(X, Z) of maps F : X ? 2Z such that for each F and
each nonempty compact subset K of X, there exists a map G ? c(K, Z) such that
G(x) ? F(x) for all x ? K. Notice the Kakutani and acyclic maps are examples of c?
maps (see [3, 4, 8] for other examples).
By a space, we mean a Hausdorff topological space. Let Q be a class of topological
spaces. A space Y is an extension space for Q (written Y ? ES(Q)) if for all X ? Q and
for all K ? X closed in X, any continuous function f0 : K ? Y extends to a continuous
function f : X ? Y .
For a subset K of a topological space X, we denote by CovX(K) the set of all
coverings of K by open sets of X (usually we write Cov(K) = CovX(K)). Let Q be a class of
topological spaces and Y a subset of a Hausdorff topological space. Given two maps
F,G : X ? 2Y and ? ? Cov(Y ), F and G are said to be ?close if for any x ? X, there
exists Ux ? ?, y ? F(x)?Ux, and w ? G(x)?Ux. A space Y is an approximate extension
space for Q (written Y ? AES(Q)) if for all ? ? Cov(Y ), for all X ? Q, for all K ? X closed
in X, and for any continuous function f0 : K ? Y , there exists a continuous function
f : X ? Y such that f K is ?close to f0.
Let X be a uniform space. Then X is Schauder admissible if for every compact subset
K of X and every covering ? ? CovX(K), there exists a continuous function (called the
Schauder projection) ?? : K ? X such that
(i) ?? and i : K X are ?close;
(ii) ??(K) is contained in a subset C ? X with C ? AES(compact).
Let X be a Hausdorff topological space and let ? ? Cov(X). X is said to be
Schauder admissible ?dominated if there exist a Schauder admissible space X? and
two continuous functions r? : X? ? X, s? : X ? X? such that r?s? : X ? X and i : X ? X
are ?close. X is said to be almost Schauder admissible dominated if X is Schauder
admissible ?dominated for each ? ? Cov(X). In [2], we established the following result.
Theorem 1.1. Let X be a uniform space and let X be almost Schauder admissible
dominated. Also suppose F ? c?(X,X) is a compact upper semicontinuous map with
closed values. Then F has a fixed point.
In our next definitions, Y will be a completely regular topological space and U an
open subset of Y .
Definition 1.2. F ? AC(U,Y ) if F : U ? K(Y ) is an acyclic compact map; here U
denotes the closure of U in Y .
Definition 1.3. F ? AC?U (U,Y ) if F ? AC(U,Y ) with x ? F(x) for x ? ?U; here ?U
denotes the boundary of U in Y .
Definition 1.4. F ? AC(Y ,Y ) if F : Y ? K(Y ) is an acyclic compact map.
Definition 1.5. If F ? AC(Y ,Y ) and p ? Y , then F {p} in AC(Y ,Y ) if there exists
an acyclic compact map R : Y ? [0,1] ? K(Y ) with R1 = F and R0 = {p} (here Rt(x) =
R(x,t)).
The following three results were established in [1]. We note that Theorem 1.7 follows
from Theorems 1.8, 1.1, and 1.6.
Theorem 1.6. Let Y be a metrizable ANR, p ? Y , and F ? AC(Y ,Y ) with F
AC(Y ,Y ). Then F has a fixed point.
{p} in
Theorem 1.7. Let Y be a completely regular topological space, U an open subset
of Y , u0 ? U, and F ? AC?U (U,Y ). Suppose there exists an acyclic compact map H :
U ?[0,1] ? K(Y ) with H1 = F, H0 = {u0}, and with x ? Ht(x) for x ? ?U and t ? (0,1).
In addition assume either of the following occurs:
(A) Y is a uniform space and Y is almost Schauder admissible dominated;
(B) Y is a metrizable ANR.
Then F has a fixed point.
Theorem 1.8. Let Y be a completely regular topological space, U an open subset
of Y , u0 ? U, and F ? AC?U (U,Y ). Suppose there exists an acyclic compact map H :
U ?[0,1] ? K(Y ) with H1 = F, H0 = {u0}, and with x ? Ht(x) for x ? ?U and t ? (0,1).
In addition, assume the following property holds:
(1.2)
(1.3)
(2.1)
(2.2)
Then F has a fixed point in U.
Let Q be a subset of a Hausdorff topological space X. Then Q is called a special retract
of X if there exists a continuous retraction r : X ? Q with r (x) ? ?Q for x ? X\Q.
Example 1.9. Let X be a Hilbert space and Q a nonempty closed convex subset of X.
Then Q is a special retract of X since we may take r (?) to be PQ(?) which is the nearest
point projection on Q.
Example 1.10. Let Q be a nonempty closed convex subset of a locally convex
topological vector space X. Then we know from Dugundji?s extension theorem that there
exists a continuous retraction r : X ? Q. If int Q = ?, then ?Q = Q so r (x) ? ?Q = Q
if x ? X. Now suppose int Q ? ?. Without loss of generality, assume 0 ? int Q. Now we
may take
x
r (x) = max 1,?(x) , x ? X,
for any G ? AC(Y ,Y ) and any p ? Y with G
in AC(Y ,Y ), G has a fixed point in Y .
{p}
there exists an acyclic compact map H : Q ?[0,1] ? K(E)
with H1 = F, H0 = u0 such that if
xj,?j j?N
here N = {1,2,... } is a sequence in ?Q ?[0,1] converging
to (x,?) with x ? H(x,?) and 0 ? ? < 1, then
H xj,?j ? Q for j sufficiently large.
where ? is the Minkowski functional on Q, that is, ?(x) = inf{? > 0 : x ? ?Q}. Note,
r (x) ? ?Q for x ? X\Q, so Q is a special retract of X.
2. Fixed point theory. In this section we present three FuriPera type theorems based
on Theorems 1.1, 1.6?1.8.
Theorem 2.1. Let E = (E,d) be a metrizable space, Q a closed subset of E, u0 ?
Q, and, Q a special retract of E. Also assume F ? AC(Q,E) with E almost admissible
dominated. In addition, suppose the following condition is satisfied:
Then F has a fixed point in Q.
Proof. Now since Q is a special retract of E, there exists a continuous retraction
r : E ? Q with r (z) ? ?Q if z ? E\Q. Consider
B = x ? E : x ? Fr (x) .
Clearly Fr : E ? K(E) is acyclic valued, upper semicontinuous, and compact. Thus Fr ?
AC(E,E), so Theorem 1.1 guarantees that B ? ?. Also since Fr is upper semicontinuous
we have that B is closed. In fact, B is compact since Fr is a compact map. It remains
to show B ? Q ? ?. To do this, we argue by contradiction. Suppose B ? Q = ?. Then
since B is compact and Q is closed, there exists a ? > 0 with dist(B,Q) > ?. Choose
m ? N = {1,2,... } with 1 < ?m. Let
1
Ui = x ? E : d(x,Q) < i
for i ? {m,m +1,...}.
(2.3)
Fix i ? {m,m + 1,...}. Now since dist(B,Q) > ?, then B ? Ui = ?. Notice also that Ui
is open, u0 ? Ui, and Fr : Ui ? K(E) is an upper semicontinuous, acyclic valued, and
compact map (i.e., Fr ? AC(Ui,E)). Let H : Q?[0,1] ? K(E) be an acyclic compact map
with H1 = F, H0 = {u0} as described in (2.1). Now let R : Ui ? [0,1] ? K(E) be given
by R(x,t) = H(r (x),t). Clearly R : Ui ? [0,1] ? K(E) is an acyclic compact map with
R1 = Fr and R0 = {u0}. Now B ? Ui = ?, together with Theorem 1.7, guarantees that
there exists
yi,?i ? ?Ui ?(0,1) with yi ? H r yi ,?i .
H r yj ,?j ? Q
for each j ? {m,m +1,...}.
D = x ? E : x ? R? r (x) for some ? ? [0,1] .
We can do this for each i ? {m,m +1,...}. Consequently,
We now look at
Now D ? ? is closed and in fact compact (so sequentially compact). This together with
1
d yj,Q = j ,
?j ? 1 for j ? {m,m +1,... }
implies that we may assume without loss of generality that
?j ? ? ? [0,1],
yj ? y ? ?Q.
In addition yj ? H(r (yj),?j) with R upper semicontinuous (so closed, [5, page 465])
guarantees that y ? H(r (y ),? ). Now if ? = 1, then y ? H(r (y ),1) = Fr (y )
which contradicts B ? Q = ?. Thus 0 ? ? < 1. But then (2.1) with xj = r (yj) ? ?Q
(note that Q is a special retract of E) and x = y = r (y ) implies {H(r (yj),?j)} ? Q
for j sufficiently large. This contradicts (2.5). Thus B ? Q ? ?, so there exists x ? Q
with x ? Fr (x) = F(x).
Remark 2.2. We can remove the assumption that Q is a special retract of E provided
we assume that
there exists a retraction r : E ? Q,
(2.4)
(2.5)
(2.6)
(2.7)
(2.8)
(2.9)
and (2.1) is replaced by the following:
there exists an acyclic compact map H : Q ? [0, 1] ? K(E)
with H1 = F, H0 = u0 such that if xj, ?j j?N
here N = {1, 2, . . . } is a sequence in Q ? [0, 1] converging
to (x, ?) with x ? H(x, ?) and 0 ? ? < 1, then
H xj, ?j ? Q for j sufficiently large.
(2.10)
(2.11)
(2.12)
(2.13)
Theorem 2.3. Let E = (E, d) be a metrizable space, Q a closed subset of E, u0 ? Q,
and Q a special retract of E. Also assume F ? AC(Q, E) with E an ANR. In addition,
assume (2.1) is satisfied and that the following condition holds:
for any G ? AC(E, E) and any p ? E, there exists
an acyclic compact map ? : E ? [0, 1] ? K(E) with
?1 = G and ?0 = {p} here ?t(x) = ?(t, x) .
Then F has a fixed point in Q.
Proof. Let r and B be as in the proof of Theorem 2.1. Notice Fr ? AC(E, E). Fix p ?
E. Now (2.11) guarantees that there exists an acyclic compact map ? : E ?[0, 1] ? K(E)
with ?1 = Fr and ?0 = {p}. This together with Theorem 1.6 guarantees that B ? ?.
Essentially the same reasoning as in Theorem 2.1 establishes the result.
Remark 2.4. In Theorem 2.3, we can replace ?Q is a special retract of E? provided
we assume (2.9) and replace (2.1) with (2.10).
Remark 2.5. From the proof of Theorem 2.3, we can see immediately that (2.11)
could be replaced by the following:
there exist p ? E and an acyclic compact map
? : E ? [0, 1] ? K(E) with ?1 = Fr and ?0 = {p}.
Our next result is a generalization of Theorem 2.3.
Theorem 2.6. Let E = (E, d) be a metrizable space, Q a closed subset of E, u0 ? Q,
and Q a special retract of E. Also assume F ? AC(Q, E) and that (2.1) and (2.11) are
satisfied. In addition, suppose the following condition holds:
E is such that for any G ? AC(E, E) and any
p ? E with G {p} in AC(E, E),
G has a fixed point.
Then F has a fixed point in Q.
Proof. Let r and B be as in the proof of Theorem 2.1. The argument in Theorem 2.3
guarantees that B ? ?. Also of course B is closed and compact. Suppose B ? Q = ?.
Then there exists a ? > 0 with dist(B, Q) > ?. Choose m ? N = {1, 2, . . . } with 1 <
?m and let Ui (i ? {m, m + 1, . . . }) be as in Theorem 2.1. Fix i ? {m, m + 1, . . .}. Note
B ? Ui = ? and Fr ? AC(Ui, E). Let H : Q ? [0, 1] ? K(E) be an acyclic compact map
with H1 = F , H0 = {u0} as described in (2.1) and let R : Ui ? [0, 1] ? K(E) be given by
R(x, t) = H(r (x), t). Clearly R : Ui ? [0, 1] ? K(E) is an acyclic compact map with R1 =
F r and R0 = {u0}. Now B ? Ui = ?, (2.13), and Theorem 1.8 guarantee that there exists
(yi, ?i) ? ?Ui ? (0, 1) with yi ? H(r (yi), ?i). We can do this for each i ? {m, m + 1, . . .}.
Consequently {H(r (yj), ?j)} ? Q for each j ? {m, m + 1, . . . }. Essentially the same
reasoning as in Theorem 2.1 from (2.5) onwards establishes the result.
Remark 2.7. In Theorem 2.6, we can replace ?Q is a special retract of E? provided
we assume (2.9) and replace (2.1) with (2.10).
Remark 2.8. In Theorem 2.6, note (2.11) could be replaced by (2.12).
Ravi P. Agarwal: Department of Mathematical Sciences, Florida Institute of Technology,
Melbourne, FL 329016975, USA
Email address:
Jewgeni H. Dshalalow: Department of Mathematical Sciences, Florida Institute of Technology,
Melbourne, FL 329016975, USA
Email address:
Advances in
Operatio ns Research
Hindawi Publishing Corporation
ht p:/ www.hindawi.com
Advances in
Journal of
Hindawi Publishing Corporation
ht p:/ www.hindawi.com
Journal of
bability
and
Statistics
Hindawi Publishing Corporation
ht p:/ www.hindawi.com
Hindawi Publishing Corporation
ht p:/ www.hindawi.com
The Scientiifc
World Journal
Hindawi Publishing Corporation
ht p:/ www.hindawi.com
International Journal of
Combinatorics
Hindawi Publishing Corporation
ht p:/ www.hindawi.com
Submit your manuscr ipts
Differential Equatio ns
Journal of
Mathematics
Hindawi Publishing Corporation
ht p:/ www.hindawi.com
En
gineering
Applied Analysis
International
Journal of
Mathematics
and
Mathematical
Sciences
Journal of
Discrete Mathematics
ht p:/ w w.hindawi.com
Journal of
Journal of
R. P. Agarwal and D. O'Regan, Homotopy and LeraySchauder principles for multi maps , Nonlinear Anal. Forum 7 ( 2002 ), no. 1 , 103  111 .
R. P. Agarwal , D. O'Regan , and S. Park , Fixed point theory for multimaps in extension type spaces , J. Korean Math. Soc . 39 ( 2002 ), no. 4 , 579  591 .
C. D. Aliprantis and K. C. Border , InfiniteDimensional Analysis , Studies in Economic Theory , vol. 4 , SpringerVerlag, Berlin, 1994 .
M. Furi and A. Vignoli , An extension of the notion of zeroepi maps to the context of topological spaces , Z. Anal. Anwendungen 21 ( 2002 ), no. 2 , 485  493 .
D. O'Regan , Fixed point theory for closed multifunctions , Arch. Math. (Brno) 34 ( 1998 ), no. 1 , 191  197 .
, Fixed point theory on extensiontype spaces and essential maps on topological spaces , Fixed Point Theory Appl . 2004 ( 2004 ), no. 1 , 13  20 .
Volume 2014 Journal of Hindawi Publishing Corporation ht p:/ www .hindawi.com