CONVERGENCE OF THE RUELLE OPERATOR FOR A FUNCTION SATISFYING BOWEN S CONDITION

Size: px
Start display at page:

Download "CONVERGENCE OF THE RUELLE OPERATOR FOR A FUNCTION SATISFYING BOWEN S CONDITION"

Transcription

1 TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 353, Number 1, Pages S (00) Article electroically published o September 13, 2000 CONVERGENCE OF THE RUELLE OPERATOR FOR A FUNCTION SATISFYING BOWEN S CONDITION PETER WALTERS Abstract. We cosider a positively expasive local homeomorphism T : X X satisfyig a weak specificatio property ad study the Ruelle operator L of a real-valued cotiuous fuctio satisfyig a property we call Bowe s coditio. We study covergece properties of the iterates L ad relate them to the theory of equilibrium states. 1. Itroductio We cosider a cotiuous map T : X X of a compact metric space ad a cotiuous fuctio : X R with some assumptios that esure the Ruelle operator L maps the space C(X; R) of real-valued cotiuous fuctios to itself ad behaves well. We obtai a covergece theorem for L as (Theorem 2.16) ad deduce results about equilibrium states. Results of this type are mostly stated whe T is a subshift of fiite type, but we use a more geeral cotext to iclude expadig maps of compact maifolds ad other examples. Let X be a compact metric space with metric d, ad let T : X X be a cotiuous surjectio. We shall assume T is positively expasive, i.e. δ 0 > 0, so that if x z 0withd(T x, T z) > δ 0. Such a umber δ 0 is called a expasive costat for T ad clearly every smaller positive umber is also a expasive costat. If we chage to a equivalet metric, the T is positively expasive i the ew metric, but the expasive costat ca chage. Reddy has show oe ca fid a equivalet metric D ad costats τ>0, > 1 such that D(x, z) <τ implies D(Tx, Tz) D(x, z). Hece T expads distaces locally i the metric D ([Re]). We also assume T is a local homeomorphism. This coditio ca be stated i several equivalet ways, which we discuss later i this sectio. The third assumptio o T is a weak specificatio coditio, which ca also be described i several ways (see Theorem 1.2). This coditio does ot ivolve periodic poits. Topologically mixig subshifts of fiite type are examples of positively expasive local homeomorphisms with the weak specificatio property, ad they are the oly subshifts with these properties. Aother importat class of examples is give by expadig differetiable maps of smooth compact coected maifolds. We have assumed T is a local homeomorphism because we wat the trasfer operators, defied by Ruelle, to map the Baach space C(X; R) of real-valued cotiuous fuctios o X, equipped with the supremum orm, to itself. For Received by the editors August 9, Mathematics Subject Classificatio. Primary 37D35; Secodary 28D20, 37A30, 37B10. Key words ad phrases. Trasfer operator, equilibrium state, etropy. 327 c 2000 America Mathematical Society

2 328 PETER WALTERS each C(X; R) the trasfer operator L : C(X; R) C(X; R) is defied by (L f)(x) = y T 1 x e(y) f(y). This is a fiite sum; sice T is positively expasive, each set T 1 x is δ 0 -separated if δ 0 is a expasive costat. Each operator L is liear, cotiuous ad positive. We take the opportuity to itroduce some otatio. We use 1 to deote the costat fuctio with value 1. The ope ball with ceter x ad radius δ will be deoted by B(x; δ). The σ-algebra of Borel subsets of X will be deoted by B(X) orbyb if o cofusio ca arise. The covex set, M(X), of all probability measures o (X, B) ca be cosidered as a subset of the dual space C(X; R) ad M(X) is compact i the weak -topology. The space of T -ivariat members of M(X) is also compact i the weak -topology ad is deoted by M(X, T). If C(X; R), the P (T,) deotes the pressure of T at ([W2]). If µ M(X), the L p µ (X) deotes the space of measurable f : X R with f p itegrable with respect to µ, p 1. The coditioal expectatio of f : X R with respect to a σ-algebra A usig µ is deoted by E µ (f/a). We sometimes write µ(f) istead of fdµ. If C(X; R), the a equilibrium state for is some µ M(X, T) with h µ (T )+µ() =P (T,)whereh µ (T ) is the etropy of the measure-preservig trasformatio T :(X, B,µ) (X, B,µ). A equivalet coditio is h σ (T )+σ() h µ (T )+µ() σ M(X, T) ([W2]). Thesymbol deotes uiform covergece. If 1ad: X R, weuse(t )(x)for 1 i=0 (T i x). Note that (L f)(x)= y T x e(t)(y) f(y). For C(X; R), 1, δ>0 defie v (, δ) =sup{ (x) (z) d(t i x, T i z) δ, 0 i 1}. If δ is a expasive costat, the v (, δ) 0as. The followig result, i which L deotes the dual of L, is well-kow, ad the first versio of it was proved by Ruelle ad is called the Ruelle operator theorem. Theorem 1.1. Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. Let C(X; R) satisfy sup v +j (T, δ) 0 as j for some δ>0 1 (ad hece for all smaller δ). There exists h C(X; R) with h>0, R with >0, ad v M(X) so that L h = h ad L ν = ν. If we ormalize h so that ν(h) =1,the f C(X; R) (L f)(x) h(x)ν(f). The spectral radius of L : C(X; R) C(X; R) is ad log = P (T,). Proof. The coditio o still holds if we chage to a equivalet metric. The result is a special case of Theorem 8 of [W1] because uder a Reddy metric D the map T satisfies the coditios of [W1]. Sice v +j (T, δ) j+ i=j v i(, δ), the coditio o is implied by =1 v (, δ) <. Oe cosequece of Theorem 1.1 is that each satisfyig the coditio of the theorem has a uique equilibrium state µ,givebyµ = h.ν (i.e. µ (f) =ν(h.f) f C(X; R)) ad T is exact with respect to µ. Also, uder the coditios of the theorem oe ca show the atural extesio of the measure-preservig trasformatio T :(X, B,µ ) (X, B,µ ) is isomorphic to a Beroulli shift ([W1]).

3 CONVERGENCE OF THE RUELLE OPERATOR 329 We wat to cosider a weaker coditio o tha the oe i Theorem 1.1. It is worth poitig out that oe eeds o assumptio o C(X; R) toget>0 ad ν M(X) withl ν = ν, as is easily see by usig the Schauder-Tychaoff fixed poit theorem o the map µ L µ (L µ)(1) of M(X). We shall cosider the followig coditio that resembles oe used by Bowe i the case of expasive homeomorphisms with a strog specificatio property ([B1]). Defiitio. We say C(X; R) satisfies Bowe s coditio (with respect to T : X X) if δ >0adC>0with the property that wheever d(t i x, T i z) δ for 0 i 1, the 1 [(T i x) (T i z)] C. i=0 Aother way to phrase this defiitio is sup 1 v (T, δ) <. If Bowe s coditio holds for some δ, the it holds for all smaller δ. Noticethatif C(X; R) satisfies the assumptio of Theorem 1.1, the it satisfies Bowe s coditio. There are examples of fuctios that satisfy Bowe s coditio but ot the assumptio of Theorem 1.1. Later results will show the Bowe coditio is a atural assumptio (Theorem 4.8). Bowe showed that whe T is a expasive homeomorphism with the strog specificatio property, the every cotiuous fuctio satisfyig his coditio has a uique equilibrium state ad this state is weak-mixig. He uses a method based o periodic poits. We cosider how much of Ruelle s operator theorem holds for C(X; R) satisfyig Bowe s coditio. We prove there is a uique ν M(X) withl ν = ν for > 0 ad this measure is tail-trivial. Recall that ν is tail-trivial meas that the oly values ν takesosetsithetailσ-algebra B = =0 T B(X) are 0 ad 1. Moreover, we show f C(X; R) (L f)(x) (L 1)(x) coverges uiformly to ν(f) (Theorem 2.16). We show has a uique equilibrium state µ ad ivestigate some properties of µ. This is doe by cosiderig measurable g-fuctios after we costruct a measurable h: X [d 1,d 2 ] (0, ) withl h = h everywhere. We use a method suggested by the work of Fa who studied the case of cotiuous g-measures for subshifts of fiite type ([F]). We also cosider covergece of L f as ad obtai a result givig L p covergece. Such a result has bee proved by Ruelle i a more geeral cotext i which the trasformatio T eed ot be a local homeomorphism ad hece the operator L does ot act o C(X; R) ([Ru]). Our measurable g-fuctio method allows us to deduce the L p covergece from a Martigale theorem. I 2 we prove the covergece theorem (Theorem 2.16) ad characterize the Bowe coditio i terms of measures. I 3 we deduce results about g-measures from the results of 2. We study equilibrium states ad the covergece of L f i 4 usig measurable g-fuctios. I 5 we cosider cotiuity properties of a measurable desity fuctio obtaied i 4. We ow discuss the assumptio that T be a local homeomorphism. The coditio ca be stated i several equivalet ways. Oe statemet, that is easier to check i examples, is that for every x X thereisaopeeighbourhoodu x of x with TU x ope such that T maps U x homeomorphically oto TU x. By a result of Eileberg ([A-H], p. 31) this is equivalet to the followig statemet i which

4 330 PETER WALTERS diam(b) deotes the diameter of the set B X: there exist δ 1 > 0, θ>0ada fuctio η :(0,δ 1 ] (0, ) with lim t 0 η(t) = 0 such that each ope subset V of X with diam(v ) <δ 1 has a decompositio of T 1 V with the followig properties: (i) T 1 V = U 1 U k for some k, whereeachu i is ope; (ii) T maps each U i homeomorphically oto V ; (iii) if i j, thed(x i,x j ) θ x i U i,x j U j ; (iv) diam(v ) < δ implies diam(u i ) < η(δ) for each i, 1 i k, ad each δ (0,δ 1 ]. If we chage to a equivalet metric, the this property still holds with differet δ 1,θ ad η. We ow show that if T : X X is a positively expasive local homeomorphism, there is some δ 2 > 0 such that wheever V is a ope subset of X with diam(v ) <δ 2 ad x V,the 1, T V is a disjoit uio y T x U y of ope sets with T mappig each U y oto V.ToseethischooseaReddymetricD with costats τ>0, >1, ad let δ 1,θ,ηcorrespod to D i the local homeomorphism coditio. Let δ satisfy δ<δ 1,δ<τad η(δ) <τ ad suppose V is ope ad diam D (V ) <δ (where the subscript shows the diameter is take for the metric D). The T 1 V = U 1 U k whereeachopesetu i has diam D (U i ) δ/ < δ. Therefore each T 1 U i ca be decomposed ito disjoit ope sets of D-diameter at most δ/ 2.By iductio we have 1 x V T V is a disjoit uio y T x U y of ope sets of D-diameter at most δ/ ad T maps each U y homeomorphically oto V icreasig distaces. If we ow revert to the origial metric d there is some δ 2 > 0 such that if V is ope ad diam(v ) <δ 2 ad x v, the 1 T V is a disjoit uio y T x U y of ope sets with T mappig each U y homeomorphically oto V. We ow cosider the coditio of weak specificatio o T. Several equivalet forms are give i the followig theorem i which B (x; ε) deotes the closed Bowe ball {y X d(t i x, T i y) ε, 0 i 1}. Theorem 1.2. For a positively expasive local homeomorphism T : X X the followig statemets are pairwise equivalet: (i) ε >0 N >0 such that x XT N x is ε-dese i X. (ii) ε > 0 M > 0 such that x, x X 1 w T (+M) x with d(t i w, T i x) ε, 0 i 1. (iii) ε >0 M >0 such that x X 1 T +M B (x; ε) =X. (iv) ε >0 M 0 such that x, x X 1, 2 1 w X with d(t i w, T i x) ε, 0 i 1 1 ad d(t 1+M+j 1 w, T j x 2 ) ε, 0 j 2 1. Proof. Note that each of the statemets is idepedet of the metric. We use B d (x; ε) for the ope ball with ceter x ad radius ε i the metric d. Assume (i) holds ad we prove (ii). Let D be a Reddy metric with D(Tx,Tx ) D(x, x ) wheever D(x, x ) < τ. Let δ 1,θ,η be associated to D by the local homeomorphism property. Let ε>0satisfyε< δ1 2, ε<τ/2adη(ε/2) <τ,ad let N correspod to ε i statemet (i) for the metric D. By statemet (i) choose y T N x with D(y,T x) <ε. By the discussio before the statemet of the theorem T B D (T x; ε) ca be writte as a disjoit uio w T T x U w of ope sets with D-diameteratmost2ε/ ad T maps each U w homeomorphically oto

5 CONVERGENCE OF THE RUELLE OPERATOR 331 B D (T x; ε). Let v = T y U x.thev T (+M) x ad for 0 i 1 D(T i xt i v ) D(T x, T v ) i <ε/ i. Hece statemet (ii) holds for the metric D, ad therefore for the origial metric d. Statemet (iii) is clearly the same as statemet (ii), ad it clearly implies statemet (iv). Now assume (iv) holds ad we prove (i). Let ε be a expasive costat ad x, z X. We wat to fid y T N x with d(z,y) <ε. By statemet (iv) M so that 1 w x with d(w,z) ε/2 add(t M+j w,t j x) ε/2 for 0 j 1. Choose a coverget subsequece w i w to get d(w, z) ε/2 ad d(t M+j w, T j x) ε/2 for all j 0. Hece T M w = x. Because of statemet (iv) we shall say that T satisfies the weak specificatio coditio if it satisfies oe, ad hece all, of the statemets i Theorem 1.2. If T satisfies the weak specificatio coditio, the T is topologically mixig. Topologically mixig subshifts of fiite type are examples of positively expasive local homeomorphisms with the weak specificatio property, ad they are the oly subshifts with these properties. Aother importat class of examples is give by expadig differetiable maps of smooth compact coected maifolds. For details about (, ε) spaig sets ad (, ε) separated sets see [W2]. Theorem 1.3. Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. The C(X; R) 1 log(l 1)(x) P (T,). Proof. By weak specificatio if ε>0 M so that x X 1 T (+M) x is (, ε) spaig, by statemet (ii) of Theorem 1.2. If { } Q (, ε) =if e (T)(x) F is (, ε) spaig, x F the ε is a expasive costat P (T,) = lim if 1 log Q (, ε). So P (T,) lim if 1 (L+M 1)(x ) = lim if 1 log(l 1)(x ) x X. Sice T is positively expasive, T x is (, ε) separated if ε is a expasive costat. If P (, ε) =sup{ x E e(t)(x) E is (, ε) separated}, thep (T,) 1 lim sup log P (, ε), ad this is actually a equality if ε is a expasive costat. Therefore P (T,) lim sup (L 1)(x ), x X. Hece 1 1 log(l 1)(x) P (T,). 2. Covergece theorem We adapt the method of Fa ([F]) to our more geeral situatio. As before, let X be a compact metric space with metric d ad let T : X X be a positively expasive, local homeomorphism with the weak specificatio property. Let C(X; R) ad let its trasfer operator be L : C(X; R) C(X; R). Let

6 332 PETER WALTERS U T : C(X; R) C(X; R) be give by (U T f)(x) = f(tx). For 1 defie P () : C(X; R) C(X; R) by (P () f)(x) = (U T L f)(x) (UT L 1)(x) = (L f)(t x) (L 1)(T x) = z T T x e(t)(z) f(z). z T T x e(t)(z) These operators have the followig properties. Theorem 2.1. For C(X; R) ad 1 let P () be defied as above. (i) P () is liear, cotiuous ad positive. (ii) P () 1=1. (iii) If l C(X; R) is so that x Xliscostat o T x,the P () () (l f) =l P f f C(X; R). (iv) L P () = L. (v) If m, thep () P (m) = P (m) P () = P (m). (vi) f C(X; R) P () f is costat o each set T x. Proof. (i), (ii), (iii) ad (vi) are clear from the defiitio. To prove (iv) we have (L P () f)(x) = e (T)(y) (P () f)(y) y T x = y T x To prove (v) let m ad the P () P (m) e (T)(y) z T x e(t)(z) f(z) z T x e(t)(z) =(L f)(x). f = P (m) f P () 1 by (iii) ad (vi) = P (m) f by (ii); P (m) P () f = U T mlm P () f UT mlm 1 = U T mlm L P () UT mlm 1 f = U m T Lm f U m T Lm 1 by (iv) = P (m) f. The dual operator P () maps M(X) itom(x) ad is cotiuous for the weak - topology. The set K () = {ν M(X) P () ν = ν} is a compact covex set which is oempty by the Schauder-Tychaoff fixed poit theorem. Note that K () = P () M(X), by Theorem 2.1 (v). By Theorem 2.1 (v) we have K (1) K (2) ad K = =1 K() is a oempty, compact covex set. Let L deote the dual of L : C(X; R) C(X; R). Theorem 2.2. Let J : M(X) M(X) be defied by J µ = L µ (L µ)(1). The P () J = J ad JM(X) K (). Proof. By iductio Jµ = (L ) µ () ((L ) µ)(1).thep Jµ = Jµ by Theorem 2 (iv). Corollary 2.3. There exist ν M(X) ad >0 with L ν = ν. Every such ν is i K.

7 CONVERGENCE OF THE RUELLE OPERATOR 333 Proof. The measures ν M(X) withl ν = ν for some >0areexactlythe fixed poits of J.SiceJ has a fixed poit by the Schauder-Tychaoff theorem, the result follows from Theorem 2.2. We shall use the followig theorem to get iformatio about P () ad about. If we let L(X; R) deote the vector space of all Borel measurable fuctios K () f : X R, the we ca cosider L as a map L : L(X; R) L(X; R)adP () as a : L(X; R) L(X; R). We shall use the followig result about a operator map P () ad later whe P is equal to L. For realvalued fuctios the expressio f f meas that for each x, f 1 (x) f 2 (x) ad f (x) f(x). P i the cases whe P is equal to P () Theorem 2.4. Let P : L(X; R) L(X; R) be a liear trasformatio which restricts to a cotiuous liear operator P : C(X; R) C(X; R). Assume P is positive i the sese that if f L(X; R) ad f 0, thepf 0. Let P act o fiite measures by fd(p µ)= Pfdµ, f C(X; R). Assume that P also has the property that wheever {f } is a sequece i L(X; R) with f f poitwise, the Pf Pf poitwise. Let µ M(X). The for all f L(X; R) with f 0 we have Pf dµ = fd(p µ). I particular, B B(X)(P µ)(b) = Pχ B dµ. Also, if f L(X; R), thef L 1 P µ (X) iff Pf L1 µ(x) ad for such f we have Pfdµ = fd(p µ). Proof. Let C = {B B(X) Pχ B dµ =(P µ)(b)}. We have X, φ Cad C is closed uder complemets ad fiite disjoit uios. We show every ope set is i C. LetU be a ope subset of X ad write it as a coutable uio j=1 B(x j; r j )of ope balls. Let C = j=1 B(x j; r j 1 )whereb(x; r) is take as empty if r 0. Each C is closed, C 1 C 2 C 3 ad =1 C = U. By Urysoh s lemma choose a cotiuous f i : X [0, 1] with χ Ci f i χ U. The f i χ U poitwise so Pf i Pχ U. By the domiated covergece theorem Pf i dµ Pχ U dµ, ad, also by the domiated covergece theorem for P µ, f i d(p µ) χ U d(p µ). Hece U C. We ext show C = B(X). Defie a measure µ 1 o X by µ 1 (B) = Pχ B dµ, B B(X). The above shows µ 1 (U) =(P µ)(u) for all ope sets U, ad sice every fiite measure o X is regular, we have µ 1 = P µ. Hece C = B(X). If f is a oegative simple fuctio, the Pf dµ = fdp µ. If f is a oegative measurable fuctio, we ca choose a sequece {f j } of oegative simple fuctios with f j f. Sice P is positive we have Pf j Pf ad the desired result follows for f, adf L 1 P µ (X) iffpf L1 µ(x). By cosiderig positive ad egative parts of f L(X; R) wehavef L 1 P µ (X) iffpf L1 µ(x) ad for such f we have Pfdµ= fd(p µ). The followig result is ow clear. Corollary 2.5. If ν K () For f L 1 ν (X) we have fdν = P () ad f L(X; R), thef L 1 ν fdν. Corollary 2.6. If ν K (), f L 1 ν (X) ad B B(X), the fdν = P () fdν. T B T B () (X) iff P f L 1 ν (X).

8 334 PETER WALTERS Proof. By Corollary 2.5 fχ T Bdν = P () (f χ T B) dν = χ T BP () fdν by Theorem 2.1 (iii). The followig characterizes the members of K (). Corollary 2.7. For ν M(X) the followig statemets are equivalet: (i) ν K (). (ii) E ν (f /T B)=P () f a.e. (v) f L 1 ν(x). (iii) E ν (f /T B)=P () f a.e. (v) f C(X; R). Proof. We have (i) (ii) by Corollary 2.6. Clearly (ii) (iii). If (iii) holds ad f C(X; R), the P () fdν = E ν (f /T B)dν = fdν, so ν K (). So the elemets of K () are those probability measures with coditioal expectatio E(f/ T B)givebyP () f. Corollary 2.8. Let ν M(X) satisfy L ν = ν for 0. If f L(X; R) ad f 0 we have L fdµ = fdν. Also, for f L(X; R) we have f L 1 ν (X) iff L f L 1 ν (X) ad for these f, L fdν = fdν. Proof. Put P equal to L i Theorem 2.4. The followig result characterizes the situatio whe K is as small as possible. Theorem 2.9. Let T : X X be a positively expasive local homeomorphism with theweakspecificatiopropertyadlet C(X; R). The followig statemets are pairwise equivalet: (i) K has oly oe member. (ii) f C(X; R) c(f) R with P () (iii) f C(X; R) c(f) R with P () f c(f). f c(f) poitwise. (iv) f C(x; R) c(f) R with L f L 1 c(f). Whe these statemets hold the uique elemet ν of K satisfies L ν = ν for some >0 ad c(f) = fdν. Proof. (i) (ii). Let K = {ν}. If (ii) holds, the c(f) = fdν by the bouded covergece theorem. If (ii) fails, the f 0 C(X; R) ad ε 0 > 0 ad sequeces j ad x j X with (P (j) f 0 )(x j ) ν(f 0 ) ε 0 j 1. We ca write this (j ) f 0 d(p δ xj ) f 0 dν ε 0 1. Choose a coverget subsequece P (j) δ xj τ M(X). Sice K () = P () M(X), we have τ K. But f 0 dτ f 0 dν ε 0 so τ ν ad this cotradicts (i). It is clear that (ii) ad (iv) are equivalet, ad that (ii) implies (iii). It remais to show (ii) implies (i). If (iii) holds ad ν K, the bouded covergece theorem gives fdν = c(f) f C(X; R) sothatν is uiquely determied.

9 CONVERGENCE OF THE RUELLE OPERATOR 335 The above proof gives the followig more geeral theorem. Let X be a compact metric space ad for each 1letP : C(X; R) C(X; R) be liear, cotiuous, positive operators with P (1) = 1 ad P P m = P max(,m). Defie K = {ν M(X) P ν = ν} ad the K 1 K 2 ad K = 1 K is oempty. Oe ca show K has oly oe member iff f C(X; R) c(f) R with P f c(f). We wat to ivestigate the extreme poits of K, ad the followig lemma is helpful. Lemma Let ν K () ad let h: X [0, ) be measurable ad hdν =1. The h ν K () iff P () h = h a.e. (v). Proof. If h ν K (),the A B(X) hdν = χ A hdν = (P () χ A ) hdν by Corollary 2.5 applied to h ν A = P () (P () χ A h)dν by Corollary 2.5 applied to ν = P () χ A P () hdν by Theorem 2.1 (iii) & (vi) = P () (χ A P () h)dν by Theorem 2.1 (iii) & (vi) = χ A P () hdν by Corollary 2.5 applied to ν = P () hdν. A Hece P () h = h a.e. (v). Coversely, if P () h = h a.e. (v), the for f C(X; R), f hdν = P () (f h)dν by Corollary 2.5 applied to ν = P () f hdν by Theorem 2.1 (iii). Therefore h ν K (). This leads to Theorem (i) The extreme poits of K areexactlytheelemetsofk that are tail-trivial. (ii) If ν 1,ν 2 are extreme poits of K, the either ν 1 = ν 2 or ν 1,ν 2 are sigular o B = =0 T B(X). Proof. (i) Let ν be tail-trivial ad ν K.Letν = pν 1 +(1 p)ν 2 with 0 <p<1 ad ν 1,ν 2 K. Sice ν 1 ν ad ν 2 ν, the above lemma gives ν 1 = h 1 ν, ν 2 = h 2 ν with h 1,h 2 both measurable (mod ν) with respect to B. Therefore h 1,h 2 are both costat (mod ν) soν 1 = ν = ν 2. Hece ν is a extreme poit of K. If ν K ad ν is ot tail-trivial, there is some B 0 B with 0 <ν(b 0 ) < 1. The χb 0 ν ν(b, χ X\B0 ν 0) ν(x\b 0) K by Lemma 2.10, ad ν = ν(b 0 ) ( χb 0 ν ν(b )+ 0) (1 ν(b 0 ))( χ X\B 0 ν ν(x\b )sothatν is ot a extreme poit of K 0).

10 336 PETER WALTERS (ii) Let ν 1,ν 2 K ad both be tail-trivial. If ν 1 ν 2,choosef 0 C(X; R) ν 1 (f 0 ) ν 2 (f 0 ). Let B i = {x X (P f 0 )(x) ν i (f 0 )}. By the Martigale theorem ν i (B i ) = 1 ([P], p. 231). Sice B 1 B 2 = we have that ν 1,ν 2 are a sigular pair. Note that B i B. We wat to show that whe satisfies Bowe s coditio there are o sigular pairs i, ad hece, by Theorem 2.11 (ii), K as oly oe member. We shall use the ope Bowe balls B (x; δ) ={y X : d(t i x, T i y) <δ,0 i 1}. For M 1adε>0lets M (ε) deote the maximum umber of poits i a (M,ε) separated set with respect to T ([W2], p. 169). For the followig lemma we oly eed T to be positively expasive. Lemma If T : X X is positively expasive ad 2δ is a expasive costat, the for M 1, 0 ad y, z X the set B (y; δ) T (+M) z has at most s M (2δ) poits. Proof. If w 1,w 2 B (y; δ) T (+M) z,thed(t i w 1,T i w 2 ) < 2δ for 0 i 1 ad T +M w 1 = T +M w 2.Sice2δ is a expasive costat, the set T (B (y; δ) T (+M) z) has the same cardiality as B (y; δ) T (+m) z,aditmustbe(m,2δ) separated. Corollary With the otatio of Lemma 2.12 {v T (+M) z B (v; δ /2 ) B (y; δ /2 ) } has at most s M (2δ) poits. Proof. Let v be i the above set. The v B (y; δ) sov T (+m) z B (y; δ), ad we ca apply Lemma The followig theorem cocers strictly positive cotiuous fuctios o X, ad we ca write such a fuctio as f = e F for F C(X; R). Theorem Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio property. Let C(X; R). For sufficietly small δ>0 M 1 with the property that 1, x, z X, F C(X; R) (L +M where A = s M (2δ)e M +v(t,δ). e F )(x) e v(f,δ) A (L +M e F )(z) Proof. Let δ > 0besosmallthat2δ is a expasive costat. By weak specificatio there exist M 1withT +M B (w; δ /4 )=X 1, w X. Hece z,y X 1 B (y; δ /2 ) T (+M) z. Let F C(X; R) ad z,x X. The y T (+M) x )(y) e F (y) e(t+m e M (2δ) ν T (+M) z )(z)+m +v(t,δ) e F (v)+v(f,δ) by associatig to each e(t+m y T (+M) x those v T (+M) z with B (y; δ /2 ) B (v; δ/2). For each y this set of v s is oempty by choice of M above, ad has at most s M (2δ) members by Corollary This gives (L +M e F )(x) e v(f,δ) A (L +M e F )(z). Corollary Let T, be as i Theorem 2.14, ad for sufficietly small δ>0 let M be give by Theorem The 1, x, z X, F C(X; R) (P +M e F )(x) e v(f,δ) A 2 +M (P e F )(z), where A = s M (2δ)e M +v(t,δ).

11 CONVERGENCE OF THE RUELLE OPERATOR 337 Proof. By Theorem 2.14 we have (L +M ad by Theorem 2.14 with F =0wehave (L +M e F )(T +M x) e v(f,δ) A (L +M e F )(T +M z), 1)(T +M z) A (L +M 1)(T +M x). Theorem Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. Let C(X; R) satisfy lim if v (T, δ) < for some δ>0, which is implied by Bowe s coditio. The: (i) K has oly oe member which is the uique ν M(X) with L ν = ν for some >0. (ii) ν is tail-trivial. (iii) f C(X; R)( L f L 1 )(x) ν(f). (iv) The uique >0 determied by (i) satisfies = L 1 dν = e P (T,) = lim ( L +1 1 L 1 ) (x) x X. Also, is the spectral radius of L : C(X; R) C(X; R). (v) If satisfies Bowe s coditio, there is a costat D>1sothat D 1 L 1(x) D 0, x X. Proof. We first use Corollary 2.15 to show K has oly oe member. let ν, ν K. If we itegrate the iequality i Corollary 2.15 i x with respect to ν, ad itegrate i z with respect to ν we get e F dν e v(f,δ) A 2 e F dν 1, F C(X; R). If A = lim if A, which is fiite by assumptio, the e F dν A 2 e F dν f C(X; R). This implies ν ν, so by Theorem 2.11 (ii) we have ν = ν. To see the above iequality gives ν ν, let C be a closed subset of X with ν (C) = 0. For ε > 0chooseaopesetU ε C with ν (U ε ) < ε, ad a Urysoh fuctio f C(X; R) withχ C f χ Uε. The τ >0 (f + τ)dν A 2 (f + τ)dν so ν(c) +τ A 2 (ν (U ε )+τ) A 2 (ε + τ), ad hece ν(c) =0. IfB B(X) has ν (B) =0,theν (C) = 0 for every closed C B ad so ν(b) =sup{ν(c) C closed, C B} = 0. Hece K has oly oe member. From Corollary 2.3 we kow there exists ν M(X) ad>0withl ν = ν, ad that such a measure is i K. Hece there is a uique such ν. The is give by = L 1 dν. Theorem 2.11 (i) gives that ν is tail-trivial. The covergece property (iii) follows by Theorem 2.9. If we put f = L 1i the covergece property we get L+1 1 L 1 (x). Hece [(L 1)(x)] 1/ ad sice L = L 1,wegetthat is the spectra radius of L : C(X; R) C(X; R). The relatio P (T,)=log follows from 1 L 1(x) log ad Theorem 1.2. To prove (v) let δ be small eough for Theorem 2.14 to hold ad let M be give by Theorem Puttig F 0 i Theorem 2.14 gives L +M 1(x) A L +M 1(z) 1, x, z X

12 338 PETER WALTERS where A =supa. If we itegrate with ν i z we get L+M 1(x) X, 1, ad if we itegrate i x we get A 1 +M A x L+M 1(z) z X, 1. If we +M let B =sup{ Li 1(x) 0 i M, x X} ad E =if{ Li 1(x) i 0 i M, x X} > i 0, the put D =max(a,b,e 1 ). The D 1 L 1(x) D x X, 0. Corollary Let T be a positively expasive local homeomorphism satisfyig the weak specificatio coditio, ad let C(X; R) satisfy Bowe s coditio. If ν, are as i Theorem 2.16 the followig statemets are pairwise equivalet: (i) h C(X; R) with h>0, L h = h, ν(h) =1ad f C(X; R) (L f)(x) h(x)ν(f). (ii) h C(X; R) with h>0 ad L h = h. (iii) (L 1)(x) coverges uiformly as. (iv) 1 1 (L i 1)(x) i=0 coverges uiformly as. i (v) { L 1 0} is a equicotiuous subset of C(X; R). (vi) { 1 1 L i 1 i=0 1} is a equicotiuous subset of C(X; R). i Proof. Clearly (i) (ii). If (ii) holds, the puttig f = h i Theorem 2.16 (iii) gives (L 1)(x) h(x) ν(h), so (iii) holds. Clearly (iii) (iv), (iv) (vi), (iii) (v) ad 1 i=0 (v) (vi). It remais to show (vi) (i). If (vi) holds, the { 1 1} i is equicotiuous ad each member is bouded from below by D 1 ad above by D, by Theorem 2.16 (v). It s closure i C(X; R) is therefore compact so that L i 1 i there is a subsequece 1 j 1 j i=0 which coverges i C(X; R) tosomeh with D 1 h. The L h = h, adν(h) = 1. Puttig f = h i Theorem 2.16 (iii) gives (L 1)(x) h(x) ad hece Theorem 2.16 (iii) gives L f hν(f). Due to Theorem 2.16 (v), part (v) of Corollary 2.17 is equivalet to ε >0 δ >0 such that wheever d(x, x ) <δ,the (L 1)(x) (L 1)(x ) 1 <ε 1. We shall deduce uiqueess of the equilibrium state of a Bowe fuctio i 4 ad also cosider covergece of L f. Oe has the followig characterizatio of the Bowe coditio. Theorem Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio property. For C(X; R) the followig are pairwise equivalet: (i) satisfies Bowe s coditio. (ii) ν M(X) ad >0 with the property that for all sufficietly small δ> 0 D δ > 1 with D 1 δ ν(b(x;δ)) D e (T)(x) δ x X, 1. (iii) ν M(X) ad >0 with the property that for some expasive costat δ 0 D δ0 > 1 with D 1 δ 0 ν(b (x; δ 0 )) e D (T)(x) δ 0 x X, 1. Proof. To show that (ii) ad (iii) are equivalet it suffices to show that if δ 1,δ 2 are expasive costats ad the property holds for δ 1, the it holds for δ 2. So assume D δ1 exists. Choose N so that d(t i x, T i y) <δ 1,0 i N 1, implies d(x, y) <δ 2 L i 1

13 CONVERGENCE OF THE RUELLE OPERATOR 339 ad choose K so that d(t i x, T i y) <δ 2,0 i K 1, implies d(x, y) <δ 1.The B +N (x; δ 1 ) B (x; δ 2 )so ( ) N 1 D 1 e e δ 2 ν(b (x; δ 2 )) x X, 1, e (T)(x) ad B (x; δ 2 ) B K (x; δ 1 )for K so ν(b(x;δ2)) D e (T)(x) δ2 (e e ) K for K. If C =sup{ ν(b(x;δ2)) x X, 1 K}, the put e (T)(x) D δ2 =max(c, D δ1 (e e ) max(k,n) ). Hece (ii) ad (iii) are equivalet. Assume (i) holds ad let δ>0 be so small that it is a expasive costat ad C =sup 1 v (T, δ) <. Let ν M(X) ad>0besothatl ν = ν. By Corollary 2.8 L fdν = fdν for f L(X; R)withf 0. By weak specificatio there is M 1withT +M B (x; δ) =X x X, 1. The L +M χ B(x;δ)(z) = e (T+M )(y) χ B(x;δ)(y) y T (+M) z has ozero terms for each z X ad e (T)(x) M C L +M χ B(x;δ)(z) s M (2δ)e T(x)+M +C by Lemma Itegratig i z gives e (T)(x) M C +M ν(b (x; δ)) e T(x) s M (2δ)e M +C x X, 1. Hece (ii) holds with D δ = s M (2δ)e M +C M. Now suppose (ii) holds ad we wish to prove (i). For sufficietly small δ ν(b(x;2δ)) ν(b (x;δ)) D 2δ D δ x X, 1. Let d(t i x, T i x ) δ /2 for 0 i 1. The x B (x ; δ) sob (x; δ) B (x ;2δ) ad e T(x) T(x ) Dδ 2 ν(b (x; δ)) ν(b (x ; δ)) D2 δ Therefore satisfies Bowe s coditio. ν(b (x ;2δ)) ν(b (x ; δ)) D 2δD 3 δ. 3. g-measures Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. Let g : X (0, 1) be cotiuous ad satisfy y T 1 x g(y) =1 x X. Such a fuctio is called a g-fuctio for T,adassigs, i a cotiuous way, a probability distributio to each of the fiite sets T 1 x. We have (L log g f)(x) = y T 1 x g(y)f(y) adl log g1 =1sothat(P () log g f)(x) = z T T x g(z) g(t 1 z)f(z). Hece P () log g = U T L() log g. By the Schauder-Tychaoff fixed poit theorem there is always at least oe µ M(X) withl log g µ = µ. Everyµ M(X) withl log g µ = µ is T -ivariat, sice if f C(X; R) f Tdµ= L log g (f T ) dµ = fdµ.ifact, usig the otatio of the previous sectio, we have, if G g = {µ M(X) L log g µ = µ} is the set of g-measures: Theorem 3.1. G g = K log g M(X, T)=K (1) log g M(X, T). Proof. We have G g K log g by Corollary 2.3 ad G g M(X, T) bytheabove. Hece G g K log g M(X, T) K (1) log g M(X, T). However, if µ K(1) log g M(X, T) the µ G g.

14 340 PETER WALTERS From Theorem 2.9 we kow that K log g has oly oe member iff f C(X; R) c(f) R with L log g f c(f). Oe ca easily show that there is a uique g-measure iff f C(X; R) c(f) R with 1 1 i=0 Li log g f c(f). Bramso ad Kalikow have give examples of g-fuctios without uique g-measures ([B-K1]). Our Theorem 2.16 gives the followig result, which was proved by Fa i the case of topologically mixig subshifts of fiite type ([F]). Theorem 3.2. Let T : X X be a positively expasive local homeomorphism with the weak specificatio coditio. Let g : X (0, 1) be cotiuous ad x X y T 1 x g(y) =1. Let log g satisfy Bowe s coditio (i.e. δ >0 ad C>0 such that d(t i x, T i x g(x)g(tx) g(t ) δ, 0 i 1 implies 1 x) g(x )g(tx ) g(t 1 x ) e C ). The f C(X; R) L log g f costat. The set K log g has oly oe member, i particular, there is a uique g-measure µ. Also, µ is exact. So for the collectio of fuctios of the form log(g) Bowe s assumptio implies the coclusio of Ruelle s theorem. There are examples of g s with L log g f costat but log g does ot satisfy Bowe s coditio ([H]). Such a example is the followig where a (0, 1). O the space X = {0, 1} Z+ let ( 3 +1 a +2) if (x 0,...,x +1 )=(1, 1,...,1, 1, 0), 0, ( 3 +1 g(x) = 1 a +2) if (x 0,...,x +1 )=(0, 1, 1,...,1, 1, 0), 0, a if x i =1 i 0, 1 a if x 0 =0adx i =1 i 1. Theorem 2.18 gives the followig Theorem 3.3. Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. For a g-fuctio g : X (0, 1) the followig are equivalet: (i) log g satisfies Bowe s coditio. (ii) µ M(X, T) with the property that for every sufficietly small δ>0 D δ > 1 with D 1 δ µ(b (x; δ)) g(x) g(t 1 x) D δ x X, 1. (iii) µ M(X, T) such that for some expasive costats δ 0 D δ0 > 1 with D 1 δ 0 µ(b (x; δ 0 )) g(x) g(t 1 x) D δ 0 x X, 0. Proof. I the proof of Theorem 2.18 we showed a measure satisfies the property i (ii) iff it satisfies the property i (iii). If (i) holds ad we choose µ with L log g µ = µ, the the proof of Theorem 2.18, with ν = µ, =1,showsµ satisfies (ii). We get (iii) implies (i) by Theorem Corollary 4.7 will give more iformatio.

15 CONVERGENCE OF THE RUELLE OPERATOR Equilibrium states I this sectio we wat to study equilibrium states ad study the covergece of L f as. We use a method ivolvig measurable g-fuctios. Let T : X X be a positively expasive local homeomorphism with the weak specificatio property. If C(X; R) satisfies Bowe s coditio, Theorem 2.16 gives a uique ν M(X) ad uique >0withL ν = ν, equals ep (T,) ad ν is tail-trivial. Also D >1withD 1 L 1(x) D x X, 1. The followig gives iformatio about solvig the equatio L h = h whe C(X; R). Theorem 4.1. Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. For C(X; R) ad >0 the followig statemets are equivalet: (i) D >1 with D 1 L 1(x) D x X, 1. (ii) measurable h: X [d 1,d 2 ] (0, ) with L h = h. (iii) h: X [d 1,d 2 ] (0, ) with L h = h. L Proof. Assume (i) holds. Let h 1 (x) = lim sup 1(x).Theh 1 : X [D 1,D] is measurable ad oe readily gets h 1 (x) (L h 1 )(x). To get h with L h = h everywhere we ca proceed as follows. We have (L h1) D L 1(x) D 2 1, x X ad sice D 1 h 1 L h 1 L2 h1 (L..., if we let h(x) = lim h1)(x) (L 2 =sup h1)(x),theh is measurable, h: X [D 1,D 2 ], ad L h = h everywhere. Hece (ii) holds. Clearly (ii) implies (iii). It remais to show (iii) implies (i). From (L h)(x) = h(x) wehaved 1 L 1(x) d 2 so that L 1(x) d2 d 1. Similarly, d 2 L 1(x) d 1 so d1 d 2 L 1(x). The discussio before Theorem 4.1 gives Corollary 4.2. If satisfies Bowe s coditio ad = e P (T,),thereexistsa measurable h: X [d 1,d 2 ] (0, ) with L h = h everywhere. From ow o we assume C(X; R) satisfies Bowe s coditio ad that ν M(X), >0, h: X [d 1,d 2 ] (0, ) aresothatl ν = ν, h is measurable, L h = h ad ν(h) = 1. The last coditio ca be attaied by replacig h i Corollary 4.2 by h ν(h). If we wat to emphasize depedece o we use ν,,h. Theorem 4.3. Let satisfy Bowe s coditio ad let ν,, h be as above. Let µ = h ν. Theµ M(X, T) ad T is a exact edomorphism with respect to µ. If h : X (0, ) is measurable with ν(h )=1ad L h = h a.e. (ν), theh = h a.e. (ν). Proof. If f C(X; R), the f T hdν = 1 L (f T h)dν = 1 f L hdν = f hdν,soµ is T -ivariat. The measures ν ad h ν are equivalet so µ is also tail-trivial, ad hece T is exact with respect to µ. If h is as i the statemet of the theorem, the µ = h ν is also i M(X, T) ad is tail-trivial ad so both µ ad µ are ergodic. Sice µ µ we must have µ = µ ad hece h = h a.e. (ν). From ow o we use µ for h ν, adwriteitµ if we wish to emphasize its relatioship to. Note L p ν(x) = L p µ(x) p 1. Also, let g = e h h T. The

16 342 PETER WALTERS g is measurable, g : X [a, b] (0, 1) ad x X y T 1 x g(y) = 1. Let L log g : L(X; R) L(X; R) be defied by (L log g f)(x) = y T 1 x g(y)f(y). We have for 1 (L 1 log gf)(x) = h(x) (L (h f))(x). From Corollary 2.8 L fdν = fdν f L 1 ν(x) sowehave L log g fdµ = fdµ f L 1 ν(x) =L 1 µ(x). We have the followig versio of a result of Ledrappier ([L]). We write B istead of B(X) adifτ M(X, T), the h τ (T ) deotes the etropy of the measurepreservig trasformatio T :(X, B,τ) (X, B,τ). Lemma 4.4. Let T : X X be a positively expasive cotiuous surjectio ad let g : X [a, b] (0, 1) be measurable ad satisfy y T 1 x g(y) =1 x X. We have h τ (T )+ log gdτ 0 τ M(X, T). Forσ M(X, T) the followig statemets are pairwise equivalet: (i) L log g fdσ = fdσ f L 1 σ(x). (ii) σ M(X, T) ad f L 1 σ(x) E σ (f /T 1 B)(x) = y T 1 Txg(y)f(y) a.e. (σ). (iii) σ M(X, T) ad h σ (T )+ log gdσ=0. Proof. We write L istead of L log g. Sice T is positively expasive, every fiite partitio of X ito sets of sufficietly small diameter is a oe-sided geerator. Hece h τ (T )=H τ (B /T 1 B) τ M(X, T). If g τ : X [0, 1] is defied a.e. (τ) by E τ (f /T 1 B)(x) = g τ (y)f(y), f L 1 τ, the Hece y T 1 Tx h τ (T )=H τ (B /T 1 B)= = h τ (T )+ log g τ dτ. y T 1 Tx log gdτ = log g/g τ dτ ( ) g 1 dτ g τ = g τ (y) y T 1 Tx =0sice y T 1 Tx g τ (y)logg τ (y) dτ(x) ( ) g(y) g τ (h) 1 dτ(x) g(y) =1 x X. Equality holds here iff log g/g τ = g/g τ 1a.e.(τ) i.e. g = g τ a.e. (τ). Hece h τ (T )+ log gdτ 0 τ M(X, T).

17 CONVERGENCE OF THE RUELLE OPERATOR 343 (i) (ii) Let f L 1 σ.the f Tdσ= L(f T ) dσ = fdσ so σ M(X, T). Also, if B B fdσ = f χ B Tdσ= L(f χ B T ) dσ T 1 B = (L(f χ B T )) Tdσ= g(y)f(y) dσ(x). T 1 B y T 1 Tx (ii) (iii) By the above if (ii) holds, the g = g σ ad (iii) holds. (iii) (i) From the above proof we kow that for τ M(X, T) h τ (T )+ log gdτ =0iffg τ = g a.e. (τ). So if (iii) holds, the g σ = g a.e. (σ). Hece for f L 1 σ Lfdσ = (Lf) Tdσ= g(y)f(y) dσ(x) = fdσ. y T 1 Tx Sice our measure µ = h ν satisfies statemet (i) of Lemma 4.4, it also satisfies (ii) ad (iii). We use this to show µ is the uique equilibrium state of. Recall that µ is a equilibrium state of if τ M(X, T) h τ (T )+ dτ h µ (T )+ dµ; equivaletly if h µ (T )+ dµ= P (T,). Theorem 4.5. Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. Let C(X; R) satisfy Bowe s coditio. The has a uique equilibrium state ad this state is µ = h ν. With respect to µt is exact. Proof. Sice log = P (T,), we kow σ M(X, T) is a equilibrium state for iff h σ (T )+ dσ =log. Sice =logg +log +logh T log h this is equivalet to h σ (T )+ log gdσ = 0. By Lemma 4.4 this is equivalet to L log g fdσ = fdσ f L 1 σ (X), which is equivalet to L fdτ = fdτ f L 1 τ(x), where τ = 1 h σ. This gives L τ = τ ad we kow, by Theorem 2.16 (i), that ν is the oly probability measure that satisfies this. Hece µ = h ν is a equilibrium state for. If σ is a equilibrium state for, theσ = c h ν for some c>0. Sice σ ad h ν are both probability measures we get c =1,adµ = h ν is the oly equilibrium state for. Exactess was proved i Theorem 4.3. Theorem 4.6. Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. Let C(X; R) satisfy Bowe s coditio, ad let = e P (T,). The followig three statemets about σ M(X, T) are pairwise equivalet: (i) σ is the uique equilibrium state µ = h ν. (ii) For every sufficietly small δ>0 E δ > 1 with E 1 δ σ(b (x; δ)) e (T)(x) E δ x X, 1. (iii) For some expasive costat δ 0 E δ0 > 1 with E 1 δ 0 σ(b (x; δ 0 )) e (T)(x) E δ 0 x X, 1.

18 344 PETER WALTERS Proof. Let ν,, h be associated to as usual. Assume (i). We have 0 < d 1 h d for some d > 1. By the proof of Theorem 2.18 we kow, ν satisfy statemet (ii) of Theorem 2.18 so we have E 1 δ µ(b (x; δ)) e (T)(x) E δ x X, 1 if E δ = dd δ. Hece (ii) holds. Clearly (ii) implies (iii). Assume (iii) holds ad we prove (i). We have 1 log σ(b (x; δ 0 )) + 1 (T )(x) log = P (T,). Itegratig with respect to σ, ad usig the Bri-Katok local etropy formula ([B-K2]), gives h σ (T )+ dσ = P (T,). This says σ is a equilibrium state for ad so must equal the uique oe µ. We have the followig special case. Corollary 4.7. Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. Let g C(X; R) be a g-fuctio so that log g satisfies Bowe s coditio. The the uique g-measure is the oly member µ of M(X, T) with either of the followig properties: (i) for every sufficietly small δ>0 E δ > 1 with E 1 δ µ(b (x; δ)) g(x)g(tx) g(t 1 x) E δ x X, 1; (ii) for some expasive costat δ 0 E δ0 > 1 with E 1 δ 0 µ(b (x; δ 0 )) g(x)g(tx) g(t 1 x) E δ 0 x X, 1. The followig geeralizes Theorem 3.3 Theorem 4.8. Let T : X X be a positively expasive local homeomorphism with the weak specificatio coditio. The followig statemets about C(X; R) are pairwise equivalet: (i) satisfies Bowe s coditio. (ii) µ M(X, T) ad >0 with the property that for all sufficietly small δ>0 E δ > 1 with E 1 δ µ(b (x; δ)) e (T)(x) E δ x X, 1. (iii) µ M(X, T) ad >0 with the property that for some expasive costat δ 0 E δ0 > 1 with E 1 δ 0 µ(b (x; δ 0 )) e E (T)(x) δ 0 x X, 1. Proof. We get (i) (ii) by Theorem 4.6. Clearly (i) (iii), ad we get (iii) (i) by Theorem 2.8. We ca ow make some deductios about equilibrium states usig a result of Quas [Q]. The result of Quas says, i our cotext, that if T : X X is a positively expasive local homeomorphism satisfyig the weak specificatio property ad µ is a ergodic member of M(X, T) with full support, the if F L µ (X) satisfies F T F = f C(X; R) a.e. there is G C(X; R) withg T G = f everywhere.

19 CONVERGENCE OF THE RUELLE OPERATOR 345 We shall let C(X; R) satisfy Bowe s coditio ad let, ν, h be so that L ν = ν, L h = h, ν(h) = 1. We kow, by Theorem 2.8, that ν(b (x, δ)) > 0 x X, 1sothatifU is ope, ν(u) > 0 because B (x, δ) U for some 1 ad some x. Hece µ = h ν is a ergodic T -ivariat measure with full support. Theorem 4.9. Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditios. Let 1, 2 C(X; R) both satisfy Bowe s coditio. The µ 1 = µ 2 iff c R ad f C(X; R) with 1 2 = c+f T f. Proof. Let µ 1 = µ 2 = µ. The g 1 = g 2 a.e. µ so e 1 h 1 = e 2 h 2 1h 1 T 2h 2 T a.e. (µ) where the i,h i correspod to i, i =1, 2. Hece 1 2 = log( 1 / 2 )+H T H a.e. (µ) whereh =logh 1 log h 2.SiceH L µ (X), the result of Quas metioed above implies f C(X; R) with 1 2 = log( 1 / 2 )+f T f everywhere. Coversely, if 1 2 = c + f T f with f C(X; R), the 1, 2 have the same equilibrium states, sice σ( 1 )=σ( 2 )+c σ M(X, T), so µ 1 = µ 2 by Theorem 4.5. This result was kow uder the stroger assumptio of Theorem 1.3 ([W1], p. 134). Theorem Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. Let C(X; R) satisfy Bowe s coditio. The equivalet statemets (i) to (vi) of Corollary 2.17 are each equivalet to: (vii) the uique equilibrium state µ of is a g-measure for a cotiuous g : X (0, 1). Proof. we show (vii) is equivalet to the existece of h C(X; R) withh>0 ad L h = h, which is statemet (ii) of Corollary If such a h exists, the g = is cotiuous ad µ is a g-measure for this g, so that (vii) holds. Now assume (vii) holds. If µ is a g-measure for a cotiuous g, theg = e h h T a.e. (µ) where L h = h ad h: X [d 1,d 2 ] (0, ) is measurable. This says log g log =logh T log h a.e. (µ), ad sice the left-had side is cotiuous ad log h L µ (X) wecausequas resulttoobtaih C(X; R) with log g log = H T H everywhere. Sice y T 1 x g(y) =1 x X we have L e H = e H. e h h T We ow tur to the covergece of L f. I the followig we use p to deote the orm i L p ν(x). The followig result was obtaied by Ruelle by aother method ([Ru]). Theorem Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. Let C(X; R) satisfy Bowe s coditio. Let >0 ad v M(X) satisfy L ν = ν ad let h: X [d 1,d 2 ] (0, ) be measurable with L h = h ad ν(h) =1. For every p 1 ad every f L p ν(x) (L f)(x) h(x)ν(f) 0 as. p Proof. L f(x) h(x)ν(f) =[L log g (f/h)(x) ν(f)]h(x) =[L log g(f/h)(x) µ(f/h)]h(x).

20 346 PETER WALTERS Sice 0 <d 1 h d 2,wehavef L p ν (X) ifff/h Lp µ (X), ad it suffices to show (L log g F )(x) µ(f ) p 0 F L p µ(x), where p ow deotes the orm i L p µ (X). By the L1 ad L p, p>1, covergece theorems ([P], pp. 231, 234) (X). Sice µ is tail-trivial, the limit is µ(f ). Hece (L log g F )(T x) µ(f ) p 0so,siceµ is T -ivariat, (L log g F )(x) µ(f ) p 0. E µ (F/T B(X)) E µ (F/ =0 T B(X)) i L p µ Corollary Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio coditio. There is a icreasig sequece of itegers 1 < 2 < 3 adasetb B(X) with µ(b) =1,whereµ is the uique equilibrium state of, such that f C(X; R) x B (L i f)(x) h(x)ν(f). i Proof. By Theorem 4.11 with f =1weget{ i } ad B B(X) withν(b) =1so that (L i 1)(x) i result. h(x) x B. Combiig this with Theorem 2.16 (iii) gives the 5. Properties of the desity fuctio h Let T : X X be a positively expasive local homeomorphism satisfyig the weak specificatio property, ad let C(x; R) satisfy Bowe s coditio. From Corollary 4.2 we kow there is a measurable h: X [d 1,d 2 ] (0, ) withl h = h ad we ca ormalize so that ν(h) =1. Let h: X [d 1,d 2 ] be defied by h(x) = lim δ 0 sup{h(y) d(x, y) <δ}. Theh is upper semi-cotiuous ad h L h. Sice both sides have the same itegral with respect to ν, we get h = L h a.e. (ν). Similarly, if h: X [d 1,d 2 ]is defied by h(x) = lim δ 0 if{h(y) d(x, y) <δ}, theh is lower semi-cotiuous ad h L h so h = L h a.e. (ν). We have h h h so that ν(h) 1 ν(h), ad by Theorem 4.3 h = a.e. (ν). h ν(h) = h h ν(h) h ν(h) Theorem 5.1. We have ν(h) everywhere. Also, { x X h(x) ν(h) = h(x) } {x X h ad h are cotiuous at x} ν(h) ad both sets are dese G δ sets with full ν-measure. Proof. Let c = ν(h) ν(h) 1. The h = ch is upper semicotiuous so that {x (h = ch)(x) < 0} is ope ad has ν-measure zero. Hece this set is empty ad ch h everywhere. We have {x X ch(x) =h(x)} = =1 { x X (h ch)(x) < 1 } so this set is a G δ set, ad sice it has ν-measure 1 it must be dese. Let x be so that ch(x) =h(x) ad we wat to show h ad h are cotiuous at x. Let ε>0. Sice h is upper semicotiuous at x ad h is lower semicotiuous at x δ >0 so that d(z,x) <δimplies h(z) < h(x) +ε ad h(x) ε<h(z). Therefore if d(z,x) <δ,the h(x) cε = c(h(x) ε) <ch(z) h(z) < h(x)+ε

21 CONVERGENCE OF THE RUELLE OPERATOR 347 so h(z) h(x) <cε. Also, d(z,x) <δimplies ( h(z) = 1+ 1 ) ε< 1c ( c h(z) 1+ 1 ) ε< 1 h(x) ε = h(x) ε<h(z) c c so that h(z) h(x) <ε. h(x) Corollary 5.2. If if x h(x) =1(i particular, if h is cotiuous at oe poit), the ν(h) =ν(h) ad ν({x X h is cotiuous at x}) =1. Proof. We have 1 ν(h) ν(h) h(x) h(x) x X so the assumptio implies ν(h) =ν(h). By Theorem 5.1 {x X h(x) =h(x)} {x X h ad h are cotiuous at x} so if h(x) =h(x), the usig h h h gives h(z) h(x) h(z) h(x) h(z) h(x) ad hece h is cotiuous at x. Therefore the desity dµ dν has a versio h ν(h) cotiuous at ν-almost every poit, ad a versio which is upper semicotiuous ad is h ν(h) which is lower semicotiuous ad is cotiuous at ν-almost every poit. The author does ot kow if there exists a cotiuous desity h whe satisfies Bowe s coditio. Refereces [A-H] N. Aoku ad K. Hiraide, Topological Theory of Dyamical Systems, North-Hollad, MR 95m:58095 [B1] R. Bowe, Some systems with uique equilibrium states, Math. Systems Theory 8 (1974), MR 53:3257 [B2] R. Bowe, Equilibrium States ad the Ergodic Theory of Aosov Diffeomorphisms, Lecture Notes i Math., vol. 470, Spriger, Berli, MR 56:1364 [B-K1] M. Bramso ad S. Kalikow, Nouiqueess i g-fuctios, IsraelJ.Math.84 (1993), MR 94h:28011 [B-K2] M. Bri ad A. Katok, O local etropy, igeometric Dyamics, Lecture Notes i Math., Vol. 1007, Spriger, Berli, MR 85c:58063 [F] Ai Hua Fa, A proof of the Ruelle operator theorem, Rev. Math. Phys. 7 (1995), MR 97e:28034 [H] F. Hofbauer, Examples for the ouiqueess of the equilibrium state, Tras. Amer. Math. Soc. 228 (1977), MR 55:8312 [L] F. Ledrappier, Pricipe variatioel et systèmes dyamiques symboliques, Z. Wahr. ud Verw. Gebiete 30 (1974), MR 53:8384 [P] K. Parthasarathy, Itroductio to Probability ad Measure, Macmilla, Lodo, MR 58:31322a [Q] A. Quas, Rigidity of cotiuous coboudaries, Bull. Lodo Math. Soc. 29 (1997), MR 99c:28054 [Re] W. Reddy, Expadig maps o compact metric spaces, Topology Appl. 13 (1982), MR 83d:54070 [Ru] D. Ruelle, Thermodyamic formalism for maps satisfyig positive expasiveess ad specificatio, Noliearity 5 (1992), MR 94a:58115 [W1] P. Walters, Ivariat measures ad equilibrium states for some mappigs which expad distaces, Tras. Amer. Math. Soc. 236 (1978), MR 57:6371 [W2] P. Walters, A Itroductio to Ergodic Theory, Graduate Texts i Math., vol. 79, Spriger, Berli, MR 84e:28017 Uiversity of Warwick, Mathematics Istitute, Covetry CV4 7AL, Eglad address: pw@maths.warwick.ac.uk

Product measures, Tonelli s and Fubini s theorems For use in MAT3400/4400, autumn 2014 Nadia S. Larsen. Version of 13 October 2014.

Product measures, Tonelli s and Fubini s theorems For use in MAT3400/4400, autumn 2014 Nadia S. Larsen. Version of 13 October 2014. Product measures, Toelli s ad Fubii s theorems For use i MAT3400/4400, autum 2014 Nadia S. Larse Versio of 13 October 2014. 1. Costructio of the product measure The purpose of these otes is to preset the

More information

Lecture Notes for Analysis Class

Lecture Notes for Analysis Class Lecture Notes for Aalysis Class Topological Spaces A topology for a set X is a collectio T of subsets of X such that: (a) X ad the empty set are i T (b) Uios of elemets of T are i T (c) Fiite itersectios

More information

Convergence of random variables. (telegram style notes) P.J.C. Spreij

Convergence of random variables. (telegram style notes) P.J.C. Spreij Covergece of radom variables (telegram style otes).j.c. Spreij this versio: September 6, 2005 Itroductio As we kow, radom variables are by defiitio measurable fuctios o some uderlyig measurable space

More information

Definition 4.2. (a) A sequence {x n } in a Banach space X is a basis for X if. unique scalars a n (x) such that x = n. a n (x) x n. (4.

Definition 4.2. (a) A sequence {x n } in a Banach space X is a basis for X if. unique scalars a n (x) such that x = n. a n (x) x n. (4. 4. BASES I BAACH SPACES 39 4. BASES I BAACH SPACES Sice a Baach space X is a vector space, it must possess a Hamel, or vector space, basis, i.e., a subset {x γ } γ Γ whose fiite liear spa is all of X ad

More information

A Proof of Birkhoff s Ergodic Theorem

A Proof of Birkhoff s Ergodic Theorem A Proof of Birkhoff s Ergodic Theorem Joseph Hora September 2, 205 Itroductio I Fall 203, I was learig the basics of ergodic theory, ad I came across this theorem. Oe of my supervisors, Athoy Quas, showed

More information

If a subset E of R contains no open interval, is it of zero measure? For instance, is the set of irrationals in [0, 1] is of measure zero?

If a subset E of R contains no open interval, is it of zero measure? For instance, is the set of irrationals in [0, 1] is of measure zero? 2 Lebesgue Measure I Chapter 1 we defied the cocept of a set of measure zero, ad we have observed that every coutable set is of measure zero. Here are some atural questios: If a subset E of R cotais a

More information

f n (x) f m (x) < ɛ/3 for all x A. By continuity of f n and f m we can find δ > 0 such that d(x, x 0 ) < δ implies that

f n (x) f m (x) < ɛ/3 for all x A. By continuity of f n and f m we can find δ > 0 such that d(x, x 0 ) < δ implies that Lecture 15 We have see that a sequece of cotiuous fuctios which is uiformly coverget produces a limit fuctio which is also cotiuous. We shall stregthe this result ow. Theorem 1 Let f : X R or (C) be a

More information

Chapter 7 Isoperimetric problem

Chapter 7 Isoperimetric problem Chapter 7 Isoperimetric problem Recall that the isoperimetric problem (see the itroductio its coectio with ido s proble) is oe of the most classical problem of a shape optimizatio. It ca be formulated

More information

Measure and Measurable Functions

Measure and Measurable Functions 3 Measure ad Measurable Fuctios 3.1 Measure o a Arbitrary σ-algebra Recall from Chapter 2 that the set M of all Lebesgue measurable sets has the followig properties: R M, E M implies E c M, E M for N implies

More information

MATH 413 FINAL EXAM. f(x) f(y) M x y. x + 1 n

MATH 413 FINAL EXAM. f(x) f(y) M x y. x + 1 n MATH 43 FINAL EXAM Math 43 fial exam, 3 May 28. The exam starts at 9: am ad you have 5 miutes. No textbooks or calculators may be used durig the exam. This exam is prited o both sides of the paper. Good

More information

Singular Continuous Measures by Michael Pejic 5/14/10

Singular Continuous Measures by Michael Pejic 5/14/10 Sigular Cotiuous Measures by Michael Peic 5/4/0 Prelimiaries Give a set X, a σ-algebra o X is a collectio of subsets of X that cotais X ad ad is closed uder complemetatio ad coutable uios hece, coutable

More information

Chapter 3. Strong convergence. 3.1 Definition of almost sure convergence

Chapter 3. Strong convergence. 3.1 Definition of almost sure convergence Chapter 3 Strog covergece As poited out i the Chapter 2, there are multiple ways to defie the otio of covergece of a sequece of radom variables. That chapter defied covergece i probability, covergece i

More information

REAL ANALYSIS II: PROBLEM SET 1 - SOLUTIONS

REAL ANALYSIS II: PROBLEM SET 1 - SOLUTIONS REAL ANALYSIS II: PROBLEM SET 1 - SOLUTIONS 18th Feb, 016 Defiitio (Lipschitz fuctio). A fuctio f : R R is said to be Lipschitz if there exists a positive real umber c such that for ay x, y i the domai

More information

Lecture 3 The Lebesgue Integral

Lecture 3 The Lebesgue Integral Lecture 3: The Lebesgue Itegral 1 of 14 Course: Theory of Probability I Term: Fall 2013 Istructor: Gorda Zitkovic Lecture 3 The Lebesgue Itegral The costructio of the itegral Uless expressly specified

More information

Sequences and Series of Functions

Sequences and Series of Functions Chapter 6 Sequeces ad Series of Fuctios 6.1. Covergece of a Sequece of Fuctios Poitwise Covergece. Defiitio 6.1. Let, for each N, fuctio f : A R be defied. If, for each x A, the sequece (f (x)) coverges

More information

University of Colorado Denver Dept. Math. & Stat. Sciences Applied Analysis Preliminary Exam 13 January 2012, 10:00 am 2:00 pm. Good luck!

University of Colorado Denver Dept. Math. & Stat. Sciences Applied Analysis Preliminary Exam 13 January 2012, 10:00 am 2:00 pm. Good luck! Uiversity of Colorado Dever Dept. Math. & Stat. Scieces Applied Aalysis Prelimiary Exam 13 Jauary 01, 10:00 am :00 pm Name: The proctor will let you read the followig coditios before the exam begis, ad

More information

Integrable Functions. { f n } is called a determining sequence for f. If f is integrable with respect to, then f d does exist as a finite real number

Integrable Functions. { f n } is called a determining sequence for f. If f is integrable with respect to, then f d does exist as a finite real number MATH 532 Itegrable Fuctios Dr. Neal, WKU We ow shall defie what it meas for a measurable fuctio to be itegrable, show that all itegral properties of simple fuctios still hold, ad the give some coditios

More information

(A sequence also can be thought of as the list of function values attained for a function f :ℵ X, where f (n) = x n for n 1.) x 1 x N +k x N +4 x 3

(A sequence also can be thought of as the list of function values attained for a function f :ℵ X, where f (n) = x n for n 1.) x 1 x N +k x N +4 x 3 MATH 337 Sequeces Dr. Neal, WKU Let X be a metric space with distace fuctio d. We shall defie the geeral cocept of sequece ad limit i a metric space, the apply the results i particular to some special

More information

Math Solutions to homework 6

Math Solutions to homework 6 Math 175 - Solutios to homework 6 Cédric De Groote November 16, 2017 Problem 1 (8.11 i the book): Let K be a compact Hermitia operator o a Hilbert space H ad let the kerel of K be {0}. Show that there

More information

Chapter 6 Infinite Series

Chapter 6 Infinite Series Chapter 6 Ifiite Series I the previous chapter we cosidered itegrals which were improper i the sese that the iterval of itegratio was ubouded. I this chapter we are goig to discuss a topic which is somewhat

More information

s = and t = with C ij = A i B j F. (i) Note that cs = M and so ca i µ(a i ) I E (cs) = = c a i µ(a i ) = ci E (s). (ii) Note that s + t = M and so

s = and t = with C ij = A i B j F. (i) Note that cs = M and so ca i µ(a i ) I E (cs) = = c a i µ(a i ) = ci E (s). (ii) Note that s + t = M and so 3 From the otes we see that the parts of Theorem 4. that cocer us are: Let s ad t be two simple o-egative F-measurable fuctios o X, F, µ ad E, F F. The i I E cs ci E s for all c R, ii I E s + t I E s +

More information

Theorem 3. A subset S of a topological space X is compact if and only if every open cover of S by open sets in X has a finite subcover.

Theorem 3. A subset S of a topological space X is compact if and only if every open cover of S by open sets in X has a finite subcover. Compactess Defiitio 1. A cover or a coverig of a topological space X is a family C of subsets of X whose uio is X. A subcover of a cover C is a subfamily of C which is a cover of X. A ope cover of X is

More information

A REMARK ON A PROBLEM OF KLEE

A REMARK ON A PROBLEM OF KLEE C O L L O Q U I U M M A T H E M A T I C U M VOL. 71 1996 NO. 1 A REMARK ON A PROBLEM OF KLEE BY N. J. K A L T O N (COLUMBIA, MISSOURI) AND N. T. P E C K (URBANA, ILLINOIS) This paper treats a property

More information

FUNDAMENTALS OF REAL ANALYSIS by

FUNDAMENTALS OF REAL ANALYSIS by FUNDAMENTALS OF REAL ANALYSIS by Doğa Çömez Backgroud: All of Math 450/1 material. Namely: basic set theory, relatios ad PMI, structure of N, Z, Q ad R, basic properties of (cotiuous ad differetiable)

More information

Boundaries and the James theorem

Boundaries and the James theorem Boudaries ad the James theorem L. Vesely 1. Itroductio The followig theorem is importat ad well kow. All spaces cosidered here are real ormed or Baach spaces. Give a ormed space X, we deote by B X ad S

More information

BIRKHOFF ERGODIC THEOREM

BIRKHOFF ERGODIC THEOREM BIRKHOFF ERGODIC THEOREM Abstract. We will give a proof of the poitwise ergodic theorem, which was first proved by Birkhoff. May improvemets have bee made sice Birkhoff s orgial proof. The versio we give

More information

A NOTE ON INVARIANT SETS OF ITERATED FUNCTION SYSTEMS

A NOTE ON INVARIANT SETS OF ITERATED FUNCTION SYSTEMS Acta Math. Hugar., 2007 DOI: 10.1007/s10474-007-7013-6 A NOTE ON INVARIANT SETS OF ITERATED FUNCTION SYSTEMS L. L. STACHÓ ad L. I. SZABÓ Bolyai Istitute, Uiversity of Szeged, Aradi vértaúk tere 1, H-6720

More information

Introductory Ergodic Theory and the Birkhoff Ergodic Theorem

Introductory Ergodic Theory and the Birkhoff Ergodic Theorem Itroductory Ergodic Theory ad the Birkhoff Ergodic Theorem James Pikerto Jauary 14, 2014 I this expositio we ll cover a itroductio to ergodic theory. Specifically, the Birkhoff Mea Theorem. Ergodic theory

More information

Notes 19 : Martingale CLT

Notes 19 : Martingale CLT Notes 9 : Martigale CLT Math 733-734: Theory of Probability Lecturer: Sebastie Roch Refereces: [Bil95, Chapter 35], [Roc, Chapter 3]. Sice we have ot ecoutered weak covergece i some time, we first recall

More information

McGill University Math 354: Honors Analysis 3 Fall 2012 Solutions to selected problems

McGill University Math 354: Honors Analysis 3 Fall 2012 Solutions to selected problems McGill Uiversity Math 354: Hoors Aalysis 3 Fall 212 Assigmet 3 Solutios to selected problems Problem 1. Lipschitz fuctios. Let Lip K be the set of all fuctios cotiuous fuctios o [, 1] satisfyig a Lipschitz

More information

MAT1026 Calculus II Basic Convergence Tests for Series

MAT1026 Calculus II Basic Convergence Tests for Series MAT026 Calculus II Basic Covergece Tests for Series Egi MERMUT 202.03.08 Dokuz Eylül Uiversity Faculty of Sciece Departmet of Mathematics İzmir/TURKEY Cotets Mootoe Covergece Theorem 2 2 Series of Real

More information

Solution. 1 Solutions of Homework 1. Sangchul Lee. October 27, Problem 1.1

Solution. 1 Solutions of Homework 1. Sangchul Lee. October 27, Problem 1.1 Solutio Sagchul Lee October 7, 017 1 Solutios of Homework 1 Problem 1.1 Let Ω,F,P) be a probability space. Show that if {A : N} F such that A := lim A exists, the PA) = lim PA ). Proof. Usig the cotiuity

More information

PRELIM PROBLEM SOLUTIONS

PRELIM PROBLEM SOLUTIONS PRELIM PROBLEM SOLUTIONS THE GRAD STUDENTS + KEN Cotets. Complex Aalysis Practice Problems 2. 2. Real Aalysis Practice Problems 2. 4 3. Algebra Practice Problems 2. 8. Complex Aalysis Practice Problems

More information

MATH301 Real Analysis (2008 Fall) Tutorial Note #7. k=1 f k (x) converges pointwise to S(x) on E if and

MATH301 Real Analysis (2008 Fall) Tutorial Note #7. k=1 f k (x) converges pointwise to S(x) on E if and MATH01 Real Aalysis (2008 Fall) Tutorial Note #7 Sequece ad Series of fuctio 1: Poitwise Covergece ad Uiform Covergece Part I: Poitwise Covergece Defiitio of poitwise covergece: A sequece of fuctios f

More information

Math 341 Lecture #31 6.5: Power Series

Math 341 Lecture #31 6.5: Power Series Math 341 Lecture #31 6.5: Power Series We ow tur our attetio to a particular kid of series of fuctios, amely, power series, f(x = a x = a 0 + a 1 x + a 2 x 2 + where a R for all N. I terms of a series

More information

The Borel hierarchy classifies subsets of the reals by their topological complexity. Another approach is to classify them by size.

The Borel hierarchy classifies subsets of the reals by their topological complexity. Another approach is to classify them by size. Lecture 7: Measure ad Category The Borel hierarchy classifies subsets of the reals by their topological complexity. Aother approach is to classify them by size. Filters ad Ideals The most commo measure

More information

BETWEEN QUASICONVEX AND CONVEX SET-VALUED MAPPINGS. 1. Introduction. Throughout the paper we denote by X a linear space and by Y a topological linear

BETWEEN QUASICONVEX AND CONVEX SET-VALUED MAPPINGS. 1. Introduction. Throughout the paper we denote by X a linear space and by Y a topological linear BETWEEN QUASICONVEX AND CONVEX SET-VALUED MAPPINGS Abstract. The aim of this paper is to give sufficiet coditios for a quasicovex setvalued mappig to be covex. I particular, we recover several kow characterizatios

More information

The Pointwise Ergodic Theorem and its Applications

The Pointwise Ergodic Theorem and its Applications The Poitwise Ergodic Theorem ad its Applicatios Itroductio Peter Oberly 11/9/2018 Algebra has homomorphisms ad topology has cotiuous maps; i these otes we explore the structure preservig maps for measure

More information

Entropy Rates and Asymptotic Equipartition

Entropy Rates and Asymptotic Equipartition Chapter 29 Etropy Rates ad Asymptotic Equipartitio Sectio 29. itroduces the etropy rate the asymptotic etropy per time-step of a stochastic process ad shows that it is well-defied; ad similarly for iformatio,

More information

ON MEAN ERGODIC CONVERGENCE IN THE CALKIN ALGEBRAS

ON MEAN ERGODIC CONVERGENCE IN THE CALKIN ALGEBRAS PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 00, Number 0, Pages 000 000 S 0002-9939(XX0000-0 ON MEAN ERGODIC CONVERGENCE IN THE CALKIN ALGEBRAS MARCH T. BOEDIHARDJO AND WILLIAM B. JOHNSON 2

More information

Lecture 3 : Random variables and their distributions

Lecture 3 : Random variables and their distributions Lecture 3 : Radom variables ad their distributios 3.1 Radom variables Let (Ω, F) ad (S, S) be two measurable spaces. A map X : Ω S is measurable or a radom variable (deoted r.v.) if X 1 (A) {ω : X(ω) A}

More information

7.1 Convergence of sequences of random variables

7.1 Convergence of sequences of random variables Chapter 7 Limit Theorems Throughout this sectio we will assume a probability space (, F, P), i which is defied a ifiite sequece of radom variables (X ) ad a radom variable X. The fact that for every ifiite

More information

Council for Innovative Research

Council for Innovative Research ABSTRACT ON ABEL CONVERGENT SERIES OF FUNCTIONS ERDAL GÜL AND MEHMET ALBAYRAK Yildiz Techical Uiversity, Departmet of Mathematics, 34210 Eseler, Istabul egul34@gmail.com mehmetalbayrak12@gmail.com I this

More information

Archimedes - numbers for counting, otherwise lengths, areas, etc. Kepler - geometry for planetary motion

Archimedes - numbers for counting, otherwise lengths, areas, etc. Kepler - geometry for planetary motion Topics i Aalysis 3460:589 Summer 007 Itroductio Ree descartes - aalysis (breaig dow) ad sythesis Sciece as models of ature : explaatory, parsimoious, predictive Most predictios require umerical values,

More information

Lecture 19: Convergence

Lecture 19: Convergence Lecture 19: Covergece Asymptotic approach I statistical aalysis or iferece, a key to the success of fidig a good procedure is beig able to fid some momets ad/or distributios of various statistics. I may

More information

lim za n n = z lim a n n.

lim za n n = z lim a n n. Lecture 6 Sequeces ad Series Defiitio 1 By a sequece i a set A, we mea a mappig f : N A. It is customary to deote a sequece f by {s } where, s := f(). A sequece {z } of (complex) umbers is said to be coverget

More information

REGULARIZATION OF CERTAIN DIVERGENT SERIES OF POLYNOMIALS

REGULARIZATION OF CERTAIN DIVERGENT SERIES OF POLYNOMIALS REGULARIZATION OF CERTAIN DIVERGENT SERIES OF POLYNOMIALS LIVIU I. NICOLAESCU ABSTRACT. We ivestigate the geeralized covergece ad sums of series of the form P at P (x, where P R[x], a R,, ad T : R[x] R[x]

More information

Homework 2. Show that if h is a bounded sesquilinear form on the Hilbert spaces X and Y, then h has the representation

Homework 2. Show that if h is a bounded sesquilinear form on the Hilbert spaces X and Y, then h has the representation omework 2 1 Let X ad Y be ilbert spaces over C The a sesquiliear form h o X Y is a mappig h : X Y C such that for all x 1, x 2, x X, y 1, y 2, y Y ad all scalars α, β C we have (a) h(x 1 + x 2, y) h(x

More information

Final Solutions. 1. (25pts) Define the following terms. Be as precise as you can.

Final Solutions. 1. (25pts) Define the following terms. Be as precise as you can. Mathematics H104 A. Ogus Fall, 004 Fial Solutios 1. (5ts) Defie the followig terms. Be as recise as you ca. (a) (3ts) A ucoutable set. A ucoutable set is a set which ca ot be ut ito bijectio with a fiite

More information

FUNDAMENTALS OF REAL ANALYSIS by. V.1. Product measures

FUNDAMENTALS OF REAL ANALYSIS by. V.1. Product measures FUNDAMENTALS OF REAL ANALSIS by Doğa Çömez V. PRODUCT MEASURE SPACES V.1. Product measures Let (, A, µ) ad (, B, ν) be two measure spaces. I this sectio we will costruct a product measure µ ν o that coicides

More information

Properties of Fuzzy Length on Fuzzy Set

Properties of Fuzzy Length on Fuzzy Set Ope Access Library Joural 206, Volume 3, e3068 ISSN Olie: 2333-972 ISSN Prit: 2333-9705 Properties of Fuzzy Legth o Fuzzy Set Jehad R Kider, Jaafar Imra Mousa Departmet of Mathematics ad Computer Applicatios,

More information

AN EXTENSION OF SIMONS INEQUALITY AND APPLICATIONS. Robert DEVILLE and Catherine FINET

AN EXTENSION OF SIMONS INEQUALITY AND APPLICATIONS. Robert DEVILLE and Catherine FINET 2001 vol. XIV, um. 1, 95-104 ISSN 1139-1138 AN EXTENSION OF SIMONS INEQUALITY AND APPLICATIONS Robert DEVILLE ad Catherie FINET Abstract This article is devoted to a extesio of Simos iequality. As a cosequece,

More information

Real Numbers R ) - LUB(B) may or may not belong to B. (Ex; B= { y: y = 1 x, - Note that A B LUB( A) LUB( B)

Real Numbers R ) - LUB(B) may or may not belong to B. (Ex; B= { y: y = 1 x, - Note that A B LUB( A) LUB( B) Real Numbers The least upper boud - Let B be ay subset of R B is bouded above if there is a k R such that x k for all x B - A real umber, k R is a uique least upper boud of B, ie k = LUB(B), if () k is

More information

Chapter 3 Inner Product Spaces. Hilbert Spaces

Chapter 3 Inner Product Spaces. Hilbert Spaces Chapter 3 Ier Product Spaces. Hilbert Spaces 3. Ier Product Spaces. Hilbert Spaces 3.- Defiitio. A ier product space is a vector space X with a ier product defied o X. A Hilbert space is a complete ier

More information

Real and Complex Analysis, 3rd Edition, W.Rudin

Real and Complex Analysis, 3rd Edition, W.Rudin Real ad Complex Aalysis, 3rd ditio, W.Rudi Chapter 6 Complex Measures Yug-Hsiag Huag 206/08/22. Let ν be a complex measure o (X, M ). If M, defie { } µ () = sup ν( j ) : N,, 2, disjoit, = j { } ν () =

More information

The log-behavior of n p(n) and n p(n)/n

The log-behavior of n p(n) and n p(n)/n Ramauja J. 44 017, 81-99 The log-behavior of p ad p/ William Y.C. Che 1 ad Ke Y. Zheg 1 Ceter for Applied Mathematics Tiaji Uiversity Tiaji 0007, P. R. Chia Ceter for Combiatorics, LPMC Nakai Uivercity

More information

Assignment 5: Solutions

Assignment 5: Solutions McGill Uiversity Departmet of Mathematics ad Statistics MATH 54 Aalysis, Fall 05 Assigmet 5: Solutios. Let y be a ubouded sequece of positive umbers satisfyig y + > y for all N. Let x be aother sequece

More information

An alternative proof of a theorem of Aldous. concerning convergence in distribution for martingales.

An alternative proof of a theorem of Aldous. concerning convergence in distribution for martingales. A alterative proof of a theorem of Aldous cocerig covergece i distributio for martigales. Maurizio Pratelli Dipartimeto di Matematica, Uiversità di Pisa. Via Buoarroti 2. I-56127 Pisa, Italy e-mail: pratelli@dm.uipi.it

More information

MASSACHUSETTS INSTITUTE OF TECHNOLOGY 6.265/15.070J Fall 2013 Lecture 21 11/27/2013

MASSACHUSETTS INSTITUTE OF TECHNOLOGY 6.265/15.070J Fall 2013 Lecture 21 11/27/2013 MASSACHUSETTS INSTITUTE OF TECHNOLOGY 6.265/15.070J Fall 2013 Lecture 21 11/27/2013 Fuctioal Law of Large Numbers. Costructio of the Wieer Measure Cotet. 1. Additioal techical results o weak covergece

More information

HOMEWORK #4 - MA 504

HOMEWORK #4 - MA 504 HOMEWORK #4 - MA 504 PAULINHO TCHATCHATCHA Chapter 2, problem 19. (a) If A ad B are disjoit closed sets i some metric space X, prove that they are separated. (b) Prove the same for disjoit ope set. (c)

More information

On Weak and Strong Convergence Theorems for a Finite Family of Nonself I-asymptotically Nonexpansive Mappings

On Weak and Strong Convergence Theorems for a Finite Family of Nonself I-asymptotically Nonexpansive Mappings Mathematica Moravica Vol. 19-2 2015, 49 64 O Weak ad Strog Covergece Theorems for a Fiite Family of Noself I-asymptotically Noexpasive Mappigs Birol Güdüz ad Sezgi Akbulut Abstract. We prove the weak ad

More information

The value of Banach limits on a certain sequence of all rational numbers in the interval (0,1) Bao Qi Feng

The value of Banach limits on a certain sequence of all rational numbers in the interval (0,1) Bao Qi Feng The value of Baach limits o a certai sequece of all ratioal umbers i the iterval 0, Bao Qi Feg Departmet of Mathematical Scieces, Ket State Uiversity, Tuscarawas, 330 Uiversity Dr. NE, New Philadelphia,

More information

Riesz-Fischer Sequences and Lower Frame Bounds

Riesz-Fischer Sequences and Lower Frame Bounds Zeitschrift für Aalysis ud ihre Aweduge Joural for Aalysis ad its Applicatios Volume 1 (00), No., 305 314 Riesz-Fischer Sequeces ad Lower Frame Bouds P. Casazza, O. Christese, S. Li ad A. Lider Abstract.

More information

Distribution of Random Samples & Limit theorems

Distribution of Random Samples & Limit theorems STAT/MATH 395 A - PROBABILITY II UW Witer Quarter 2017 Néhémy Lim Distributio of Radom Samples & Limit theorems 1 Distributio of i.i.d. Samples Motivatig example. Assume that the goal of a study is to

More information

Metric Space Properties

Metric Space Properties Metric Space Properties Math 40 Fial Project Preseted by: Michael Brow, Alex Cordova, ad Alyssa Sachez We have already poited out ad will recogize throughout this book the importace of compact sets. All

More information

Topologie. Musterlösungen

Topologie. Musterlösungen Fakultät für Mathematik Sommersemester 2018 Marius Hiemisch Topologie Musterlösuge Aufgabe (Beispiel 1.2.h aus Vorlesug). Es sei X eie Mege ud R Abb(X, R) eie Uteralgebra, d.h. {kostate Abbilduge} R ud

More information

Empirical Processes: Glivenko Cantelli Theorems

Empirical Processes: Glivenko Cantelli Theorems Empirical Processes: Gliveko Catelli Theorems Mouliath Baerjee Jue 6, 200 Gliveko Catelli classes of fuctios The reader is referred to Chapter.6 of Weller s Torgo otes, Chapter??? of VDVW ad Chapter 8.3

More information

Research Article A Note on Ergodicity of Systems with the Asymptotic Average Shadowing Property

Research Article A Note on Ergodicity of Systems with the Asymptotic Average Shadowing Property Discrete Dyamics i Nature ad Society Volume 2011, Article ID 360583, 6 pages doi:10.1155/2011/360583 Research Article A Note o Ergodicity of Systems with the Asymptotic Average Shadowig Property Risog

More information

Seunghee Ye Ma 8: Week 5 Oct 28

Seunghee Ye Ma 8: Week 5 Oct 28 Week 5 Summary I Sectio, we go over the Mea Value Theorem ad its applicatios. I Sectio 2, we will recap what we have covered so far this term. Topics Page Mea Value Theorem. Applicatios of the Mea Value

More information

SOME SEQUENCE SPACES DEFINED BY ORLICZ FUNCTIONS

SOME SEQUENCE SPACES DEFINED BY ORLICZ FUNCTIONS ARCHIVU ATHEATICU BRNO Tomus 40 2004, 33 40 SOE SEQUENCE SPACES DEFINED BY ORLICZ FUNCTIONS E. SAVAŞ AND R. SAVAŞ Abstract. I this paper we itroduce a ew cocept of λ-strog covergece with respect to a Orlicz

More information

The Boolean Ring of Intervals

The Boolean Ring of Intervals MATH 532 Lebesgue Measure Dr. Neal, WKU We ow shall apply the results obtaied about outer measure to the legth measure o the real lie. Throughout, our space X will be the set of real umbers R. Whe ecessary,

More information

5. Matrix exponentials and Von Neumann s theorem The matrix exponential. For an n n matrix X we define

5. Matrix exponentials and Von Neumann s theorem The matrix exponential. For an n n matrix X we define 5. Matrix expoetials ad Vo Neuma s theorem 5.1. The matrix expoetial. For a matrix X we defie e X = exp X = I + X + X2 2! +... = 0 X!. We assume that the etries are complex so that exp is well defied o

More information

Chapter IV Integration Theory

Chapter IV Integration Theory Chapter IV Itegratio Theory Lectures 32-33 1. Costructio of the itegral I this sectio we costruct the abstract itegral. As a matter of termiology, we defie a measure space as beig a triple (, A, µ), where

More information

MAS111 Convergence and Continuity

MAS111 Convergence and Continuity MAS Covergece ad Cotiuity Key Objectives At the ed of the course, studets should kow the followig topics ad be able to apply the basic priciples ad theorems therei to solvig various problems cocerig covergece

More information

62. Power series Definition 16. (Power series) Given a sequence {c n }, the series. c n x n = c 0 + c 1 x + c 2 x 2 + c 3 x 3 +

62. Power series Definition 16. (Power series) Given a sequence {c n }, the series. c n x n = c 0 + c 1 x + c 2 x 2 + c 3 x 3 + 62. Power series Defiitio 16. (Power series) Give a sequece {c }, the series c x = c 0 + c 1 x + c 2 x 2 + c 3 x 3 + is called a power series i the variable x. The umbers c are called the coefficiets of

More information

Research Article Approximate Riesz Algebra-Valued Derivations

Research Article Approximate Riesz Algebra-Valued Derivations Abstract ad Applied Aalysis Volume 2012, Article ID 240258, 5 pages doi:10.1155/2012/240258 Research Article Approximate Riesz Algebra-Valued Derivatios Faruk Polat Departmet of Mathematics, Faculty of

More information

The Choquet Integral with Respect to Fuzzy-Valued Set Functions

The Choquet Integral with Respect to Fuzzy-Valued Set Functions The Choquet Itegral with Respect to Fuzzy-Valued Set Fuctios Weiwei Zhag Abstract The Choquet itegral with respect to real-valued oadditive set fuctios, such as siged efficiecy measures, has bee used i

More information

Approximation by Superpositions of a Sigmoidal Function

Approximation by Superpositions of a Sigmoidal Function Zeitschrift für Aalysis ud ihre Aweduge Joural for Aalysis ad its Applicatios Volume 22 (2003, No. 2, 463 470 Approximatio by Superpositios of a Sigmoidal Fuctio G. Lewicki ad G. Mario Abstract. We geeralize

More information

Advanced Stochastic Processes.

Advanced Stochastic Processes. Advaced Stochastic Processes. David Gamarik LECTURE 2 Radom variables ad measurable fuctios. Strog Law of Large Numbers (SLLN). Scary stuff cotiued... Outlie of Lecture Radom variables ad measurable fuctios.

More information

ANSWERS TO MIDTERM EXAM # 2

ANSWERS TO MIDTERM EXAM # 2 MATH 03, FALL 003 ANSWERS TO MIDTERM EXAM # PENN STATE UNIVERSITY Problem 1 (18 pts). State ad prove the Itermediate Value Theorem. Solutio See class otes or Theorem 5.6.1 from our textbook. Problem (18

More information

n p (Ω). This means that the

n p (Ω). This means that the Sobolev s Iequality, Poicaré Iequality ad Compactess I. Sobolev iequality ad Sobolev Embeddig Theorems Theorem (Sobolev s embeddig theorem). Give the bouded, ope set R with 3 ad p

More information

Lecture 8: Convergence of transformations and law of large numbers

Lecture 8: Convergence of transformations and law of large numbers Lecture 8: Covergece of trasformatios ad law of large umbers Trasformatio ad covergece Trasformatio is a importat tool i statistics. If X coverges to X i some sese, we ofte eed to check whether g(x ) coverges

More information

An Introduction to Randomized Algorithms

An Introduction to Randomized Algorithms A Itroductio to Radomized Algorithms The focus of this lecture is to study a radomized algorithm for quick sort, aalyze it usig probabilistic recurrece relatios, ad also provide more geeral tools for aalysis

More information

Introduction to Optimization Techniques

Introduction to Optimization Techniques Itroductio to Optimizatio Techiques Basic Cocepts of Aalysis - Real Aalysis, Fuctioal Aalysis 1 Basic Cocepts of Aalysis Liear Vector Spaces Defiitio: A vector space X is a set of elemets called vectors

More information

Lecture 7: Properties of Random Samples

Lecture 7: Properties of Random Samples Lecture 7: Properties of Radom Samples 1 Cotiued From Last Class Theorem 1.1. Let X 1, X,...X be a radom sample from a populatio with mea µ ad variace σ

More information

On forward improvement iteration for stopping problems

On forward improvement iteration for stopping problems O forward improvemet iteratio for stoppig problems Mathematical Istitute, Uiversity of Kiel, Ludewig-Mey-Str. 4, D-24098 Kiel, Germay irle@math.ui-iel.de Albrecht Irle Abstract. We cosider the optimal

More information

AN ARC-LIKE CONTINUUM THAT ADMITS A HOMEOMORPHISM WITH ENTROPY FOR ANY GIVEN VALUE

AN ARC-LIKE CONTINUUM THAT ADMITS A HOMEOMORPHISM WITH ENTROPY FOR ANY GIVEN VALUE AN ARC-LIKE CONTINUUM THAT ADMITS A HOMEOMORPHISM WITH ENTROPY FOR ANY GIVEN VALUE CHRISTOPHER MOURON Abstract. A arc-like cotiuum X is costructed with the followig properties: () for every ɛ [0, ] there

More information

7.1 Convergence of sequences of random variables

7.1 Convergence of sequences of random variables Chapter 7 Limit theorems Throughout this sectio we will assume a probability space (Ω, F, P), i which is defied a ifiite sequece of radom variables (X ) ad a radom variable X. The fact that for every ifiite

More information

1 Convergence in Probability and the Weak Law of Large Numbers

1 Convergence in Probability and the Weak Law of Large Numbers 36-752 Advaced Probability Overview Sprig 2018 8. Covergece Cocepts: i Probability, i L p ad Almost Surely Istructor: Alessadro Rialdo Associated readig: Sec 2.4, 2.5, ad 4.11 of Ash ad Doléas-Dade; Sec

More information

6.3 Testing Series With Positive Terms

6.3 Testing Series With Positive Terms 6.3. TESTING SERIES WITH POSITIVE TERMS 307 6.3 Testig Series With Positive Terms 6.3. Review of what is kow up to ow I theory, testig a series a i for covergece amouts to fidig the i= sequece of partial

More information

It is always the case that unions, intersections, complements, and set differences are preserved by the inverse image of a function.

It is always the case that unions, intersections, complements, and set differences are preserved by the inverse image of a function. MATH 532 Measurable Fuctios Dr. Neal, WKU Throughout, let ( X, F, µ) be a measure space ad let (!, F, P ) deote the special case of a probability space. We shall ow begi to study real-valued fuctios defied

More information

2 Banach spaces and Hilbert spaces

2 Banach spaces and Hilbert spaces 2 Baach spaces ad Hilbert spaces Tryig to do aalysis i the ratioal umbers is difficult for example cosider the set {x Q : x 2 2}. This set is o-empty ad bouded above but does ot have a least upper boud

More information

5 Birkhoff s Ergodic Theorem

5 Birkhoff s Ergodic Theorem 5 Birkhoff s Ergodic Theorem Amog the most useful of the various geeralizatios of KolmogorovâĂŹs strog law of large umbers are the ergodic theorems of Birkhoff ad Kigma, which exted the validity of the

More information

Fall 2013 MTH431/531 Real analysis Section Notes

Fall 2013 MTH431/531 Real analysis Section Notes Fall 013 MTH431/531 Real aalysis Sectio 8.1-8. Notes Yi Su 013.11.1 1. Defiitio of uiform covergece. We look at a sequece of fuctios f (x) ad study the coverget property. Notice we have two parameters

More information

MAXIMAL INEQUALITIES AND STRONG LAW OF LARGE NUMBERS FOR AANA SEQUENCES

MAXIMAL INEQUALITIES AND STRONG LAW OF LARGE NUMBERS FOR AANA SEQUENCES Commu Korea Math Soc 26 20, No, pp 5 6 DOI 0434/CKMS20265 MAXIMAL INEQUALITIES AND STRONG LAW OF LARGE NUMBERS FOR AANA SEQUENCES Wag Xueju, Hu Shuhe, Li Xiaoqi, ad Yag Wezhi Abstract Let {X, } be a sequece

More information

Notes #3 Sequences Limit Theorems Monotone and Subsequences Bolzano-WeierstraßTheorem Limsup & Liminf of Sequences Cauchy Sequences and Completeness

Notes #3 Sequences Limit Theorems Monotone and Subsequences Bolzano-WeierstraßTheorem Limsup & Liminf of Sequences Cauchy Sequences and Completeness Notes #3 Sequeces Limit Theorems Mootoe ad Subsequeces Bolzao-WeierstraßTheorem Limsup & Limif of Sequeces Cauchy Sequeces ad Completeess This sectio of otes focuses o some of the basics of sequeces of

More information

DANIELL AND RIEMANN INTEGRABILITY

DANIELL AND RIEMANN INTEGRABILITY DANIELL AND RIEMANN INTEGRABILITY ILEANA BUCUR We itroduce the otio of Riema itegrable fuctio with respect to a Daiell itegral ad prove the approximatio theorem of such fuctios by a mootoe sequece of Jorda

More information

ABOUT CHAOS AND SENSITIVITY IN TOPOLOGICAL DYNAMICS

ABOUT CHAOS AND SENSITIVITY IN TOPOLOGICAL DYNAMICS ABOUT CHAOS AND SENSITIVITY IN TOPOLOGICAL DYNAMICS EDUARD KONTOROVICH Abstract. I this work we uify ad geeralize some results about chaos ad sesitivity. Date: March 1, 005. 1 1. Symbolic Dyamics Defiitio

More information

Application to Random Graphs

Application to Random Graphs A Applicatio to Radom Graphs Brachig processes have a umber of iterestig ad importat applicatios. We shall cosider oe of the most famous of them, the Erdős-Réyi radom graph theory. 1 Defiitio A.1. Let

More information

Math 220A Fall 2007 Homework #2. Will Garner A

Math 220A Fall 2007 Homework #2. Will Garner A Math 0A Fall 007 Homewor # Will Garer Pg 3 #: Show that {cis : a o-egative iteger} is dese i T = {z œ : z = }. For which values of q is {cis(q): a o-egative iteger} dese i T? To show that {cis : a o-egative

More information