Ergodic properties of Fleming-Viot processes with selection
Original Paper
Ergodic properties of Fleming-Viot Processes with selection
Seiichi ITATSU*
Department of Mathematics, Faculty of Science, Shizuolm University, Ohya 836, Shizuol<a 422-8529
(Received November 30, 2001 )
Abstract : Under the condition that mutation process is uniforrnly ergodic,
it is shown that for small o the Fleming-Viot process with selection o is uni- formly ergodic. In particular if the distribution of mutation process converges exponentially fast , then the the distribution of the Fleming-Viot process with no selection converges exponentially fast with the same exponent.
2000 Mathematical Subject Classification:60J70
Key words : measure valued process, semigroup, ergodicity
1. Introduction of Fleming-Viot processes wiph selection Let us denote the operator L of the infinitesimal generator in C(R*)
defined by the following:
L - *,D,pt(66, - p ) Lrr*.
E,I %ipt. f p i(or,ipt ->ro nenp)) h
o i,i:L
This defines the infinitesimal generator of a Markov process on A6 : {p -
(pt,- - -,prc) : pr ) 0,...,prc ) O,pt+ -. - * prc - 1), this process is called
the Wright-Fisher diffusion model according to Ethier and Kurt, l4l. Here pe is a gene frequency of type i. According to Ethier and Kurtz [3], this operator can be generalized as follows: Let E be a compact metric space and
* Bmail: [email protected]
Se五ch口園肛
SU
P (E) be the space of all probability measures on E. Let us denote (f , lr) -
[p fdp. Let A be the generator for a Feller semigroup {"(t)} or C(E). Fbr uV fr,.-.,f* €D(A) and F e C2(R*) let e\.i - F((fr,F),...,(f*,ti)
: F((f,p)) and let us denote
Lvii 13'|tr+.,F)-Ua,tr)( o"
- 2 L__r((fufi, ri - ffn, ti1i, rl)&((f, p))
(1) + if<ot,,p)+(Bfn,t'))ffUr,rll
+ itfi(fu " r)o, t') - (fa, trl(o, t'l\fft<t, rll
i:I
Here E is the space of genetic types and A is a mutation operator on C (E) which is the generator for a Markov process in E , B is a recombination operator from C(E) to C(82), and o : o(r,u) is a symmetric function on E x E which is selection parameters for types r,y € E, md zr is a
projection defined by r(r,U): r . According to [3], this operator defines a generator corresponding to a Markov process on P (E) in the sense that the Cp@) [0, *) martingale problem for .C is well posed. This process is called the Fleming-Viot process. In this paper we consider in the case B - 0 the
formula
Lvii - ; ,f_"(fuf i, tr) - (ft, riui, ri)ffi((r,p))
o i,i:l
+Σ
% (4ん,μ)民を(ば,μ
))i:l
+ it ,i,:l ($r o n)o, p2) - Ur,, tib, t'llf*f<r, rll.
We say the semigroup t"(r)) is uniformly ergodic if for some stationary
distribution u, ll"(t) - (., u)Lll + 0 as t + oo where ll . ll is the uniform operator norm. In the case that the mutation semigroup is uniformly ergodic, we consider the uniformly ergodic properties of the semigroup corresponditrg
to generators L in the form (2) and consider the order of convergence. We can obtain a uniformly ergodic theorem if selection parameters are small.
In [5] the ergodic theorerns are proved in the case that the coupling in the
(2)
Ergodic propertieS ofFleming‐ Vlot processes v7it selection 3
mutation process holds,which cont〔 狙ns the case that the mutation operator スis of the fo...1
(3) スノ(″
)==│:/(∫ (ξ
)……ノ(″))ν
(dξ)with θ>O and a(五
stribution
ν on」『 . On the other hand in the case B==0
it is proved in[l thtt the reversibihty of the process is holds only ifス
is ofthe昴ove form(3).
ALCk■ OWledgement l would hke to thank Professor Steward N.Ethier for helpful suggestions.I would like to th〔
狙k Professor Tokllzo Shigal for mally valuable(五scussions.
2。 An ergodic theorem in the case without selection
lt is known when the mutttion operttor is of the form(3),trttSition
function converges exponentia■y fast as follows
Theorem l。 (Ether and G五 蹴 hs μ],Ethier ttd Kurtz pl,shiga pl,LⅦ
6
p])Lcι スb̀α
η"cmι
θ
r as
ス∫(″
)`
=二 ::プ[(∫(ξ)¨―ノ(″))ノ
(αξ),ιん
̀η
l九c c"づsts a sια οηαη αづ
stl%bttι
づοη Hθ,ν
sttcん J力αι ιれct%ηsづ οη prob―α bづ Jづ ιν P(t,μ
,0)げ流
C SC鶴づ
g"Ψ {鶴(t)}SatづJCS ttαι
‖
P(t,μ ,。 )一
Πθ,ν llυ
αr≦1‑4〈 t),
where‖
・‖υarづS tο
ιαJ
υαttαιづοη αηα α。(t)SaιづJCS ttαιeXp(― λ
lt)≦
1‑αo(t)≦
(1+θ)exp(一 人lt) 磁c
λl=:.We can obtain a generalizёd ergodic theorem.In this section we consider thtt E is a locally compact,separable metric space and th就 スis the gener―
就Or of a Fener semigroup{T(t)}On the space σ
(E)of COntinuous hnctions
on E withマ劉mshng tt inhity.Let us consider the Fleming―
Viot processdeflned by the ge■
erator of the form(2)with σ=0・
In this case demte theSeiich nlttU
generttOr£ by£。and by pl the wellpOsedness Ofthe martingde prOblem fOr
£。has been prOved and the ergodic theorerrl has been prOved in the sense of weak cOIlvergence under the cOndition that the lnutation operator is ergOdic in the sense of wealkly cOnvergenceo Let us denote the serrugroup of the gen̲
erttor£。by{γ(t)},then we ha℃ th就
Theorem 2。 Zcι λl>0。 ■ んο
Jα
s ttc cοηαづοη(A):ι
ん̀
̀蒻st a cθηsι
αηι 埼 >O αηα α sιαιづοηαηttsι
れれιづοη tt sttCん 流αιルr αην∫∈0(E)α ηα αην ι>0
‖
T(ι)∫ ―
(∫,均)1‖ ≦』 島
exp(―λ
lサ)││∫││, ゲα η ごθ η :ν ゲづ ιん ο
Jお流
c cοη鹿 θ η (B):ι ん
̀ "づ
sι
a cοη sι α ηι
ν
>Oα ηご
a stα
ι づ θ η α
rνα づ sι
ttb鶴θ η Ⅱ
o sαι 亀ル
jη′流α ιル r α η ν ψ∈σ
(2(E))αη α
α ην ι >0
‖γ(t)ψ 一(9,Π。)111≦ ″
rexp(―
λlサ)│1911。For the proOfthe next TheOrem win be used.Let us denOte{μ ι
}the Fleming̲
Viot process cOrresponding tO£ 。。
Theorem 3(Ether ttd Kllrtz pl).Zθι S=Σ鷹10(Eり bC α Ψαccげαづct
S鶴
鶴 げ βαηαcL spaccs αηJごヴ ηθα Mαttθυ
p"ccss
θη S ttιん流c θcη
cmιοr
£F(ノ)==(:)(F(【
:)1≦忍 ≦た
゛
lF)ノ
)一 F(∫))+!:BF(7L(サ)∫
)‐―F(ノ)ル
r∫
∈∂(Eり υんc
(Φ才)ノ
)(″ 1,・
…,″
ん̲1)=∫(″1,… 。,物
̲1,″り,″ J,・
…,″ん‑1)ルrた ≧2 αηαl≦ づ
<ブ
≦たαηαノ∈0(Eん)α
ηα■(サ)=T(ι)Θ
た。
∂
(」
」F;づ
:;I:1:ll)'こ (2}メ:メ(1}‖│;IΨ、:;;亀′場
;J子
樗i411'鶴′づ
・ 9 SCηsc. Jy(t)∈
whereY(O) -f.
易ι[(∫
,μ
夕)]=EI(y(サ),μⅣ(t))]Ergodic properties of Fleming‐
Vlot processes wim sdection 5 P知げ げ 動 Cο 鶴 2。Wc assulne the condtioll(A).Then fOr allyノ
≧ 1狙
dg∈
0(EJ)(4)‖TJ・
(t)g― (g,イ
)││≦ ノ‰eXp(一人lt)│lglト
Letた
≧ 1,and∫ ∈ 0(Eた),狙d let毎
̲J=inf{t>QⅣ (t)=J}fOrブ =1,・ …,た ‑l and η
=∞
,then by the茄ove theorem
Eμ(ノ
,μ
夕)=EI(y(t),μⅣ(t))]=Σ EI(y(t),μん→+1);η ̲1≦
t<η].J=1
Here
Ⅳ(t)is a detth process,which julnpsた → た‑1,with rtteた(た ‑1)/2forた
≧2 alld η―η̲l htt expone戴
hl出st五butim with pttattter(た 5J)。
By[ll it h01ds thtt P(%̲1>t)≦ 1‑α:(t)証 ld C t≦ 1‑α
:(t)≦ 3c t.Let
5° =音 ΣをくゴΦ絆・Lr
η̲1≦
t<η Wehaκy(ι)=写̲J+1(t η̲1)5。 Jttaη
̲J+2(η ‑1 η
̲2)・ …5°と耽(71)∫・Next p乱 an integerづ >O such thtt λ
l<Cl),then We can see forた >づOL頑 硼 看 1≦ 』 1<鳴
arld
(6)P17t̲づ >J≦ Elexp(λl(■̲づ
一
t));η卜 を
>」≦
Cttλlι
Elexp(λ l%̲劇・ Let
Zo=(y(■
̲̀),イ ),
碗(t)=y(ι)̲Zol,
ZF・ =(時(■
̲j→
),ガ打′),
Ч汁1(サ)=場(ι)一 ZJ・
1,
ノ=1,…,づ ‑1,and v」e haК ̀
E(y(ι),μ
Ⅳ← ))=Σ
E[(y(ι),μを
ゴ+1);η +た
̲o̲1≦ t<η+た ̲り
]+J=1
EI(y(t),μlV(1));η
̲を
>t]Se五
ch
ΠMttSU
0 ′
=ΣEK場
(t)+Σ
る‑11,ノ J+1);η十臓
‑1≦
t<η十か』J=l J=1
十
E[(y(ι
),μⅣ(t));η卜づ>司=ΣEI(yJ・
(サ
),μ̀丁
J+1);η
+ん̲̲̀1≦ t<ηttλ̲』
J=1
̀ J
+Σ
〕】EEI(ZJ̲11,μ
̀→+1);η
+た ̲り ̲1≦
サ<η
十た一̀]+EI(y(t),μ
Ⅳ
(ι
));η卜 を>tl
J=lJ=1=ΣE[ 為
(サ
),μうJ+1);η
十た一う
‑1≦
t<η十た一』J=1
+IΣ
Σ
E[ZJ̲1;η+た ̲を ‑1≦
t<η十た ―り ]十
E[(y(t),μN(ι));ηトゥ
>サ].
J=lJ=J
(7)E(y(ι),μⅣ
O)=Σ
E[乃
(ι
),μ̀ J+1);η
十ん―リー1≦t<η
十ん―J
J=1
+IΣ
E[ろ
̲1;7J十た‑0‑1≦ti+EI(y(サ
),μⅣ(̀));η卜を>ι]・
J=1
Because
(8) yJ.(ι)
=(島
一J+1(t η
卜 を十J‑1) (0,ス
プ+1)1)0(づ→
+2)(z.+2(η
十た一
̀‑1 η
ttλ―̀‑2) (°,スプ
+2)1)
・…3(̀)(z(η̲を+1‑■―
̀) (0,ス )1)y(.̲̀)
fOr η十ん―を
‑1≦
ι<η十ん―り,from(4)we haК‖yJ・(t)││≦ づ
(づ
‑1)・ …(づ
―ノ+1)ルイexp(一人1(t― η卜を))‖∫││
狙d
lZJ≦ ‖時(η十た二
̀)‖
・
Ergodic properties ofFleming‐
Vlot processes宙 ぬselection 7 So by(5),(6),and(7)│コ
μ
[(∫,μ :)]―
EI(y(■‑1),ぬ)]│≦
)EIE[に
為(ι
),μ̀ J+1);η +た
̲̲̀1≦t<η
十た一̀ll+Σ
IE[ZJ;η卜̀十J>JI J=l J=Ю
十
IEI(y(t),μ
Ⅳ(ι
));η卜ぅ>ι
]│
≦ Σ〕づ
(づ
‑1)。 …(づ
―ノ+1)ル昭Elexp(一人1(t―
■■));η+卜
を‑1≦
ι<η +た ̲り
]││∫‖J=1
,‑1+Σ
づ (づ
‑1)・… (づ ―
J+lMElexp(一人
1(7J十ん―う
η卜 り
));7J十
ん―づ
>t]││ノ‖
J=1+2P[η̲づ >JII∫‖
≦
νeXp(一人
lt)││ノ
‖where
ν ≪
2Σ呻 →坤抑 瑚弼→
J里1
づ狙d ν
depend only on
λl.Becttse∪ん>り {9(μ )=(∫ ,μ
り:∫ ∈0(Eん)}is densc in
σ(2(E))by the mesz'represent乱
lon theOrem the condition(B)holds.
Collversdy we assume the coFldition(B)hddS・ There odsts a stttiorLary
distribution
Ⅱ。証ld collstants y,λ
>O sttisfying th乱‖γ(ι)9‑(り,Πo)111≦ νexp(―λlt)‖91ト
ヽ
Let 9(μ)=(∫,μ
)・
Then γ(t)ψ(μ)=(T(ι)ノ
,μ),let均 ∈2(E)by(∫,埼
)=/〈 ∫ ,μ
)ΠO(中
)so that
(9,Πo)=(ノ,埼
)・
Therefore‖
γ
(t)9‑(9,Ⅱo)111=‖(T(サ)∫ ,。
)一 (ノ,均)1‖
=‖T(ι)∫ ―
(∫,均)111。By‖
9‖=‖ ∫‖
,the inequality of(A)holdS・QoEoD。
3. Ergodic theorenrs with selection
8 seiich FrASU
In ths section we assllme thtt Z is cOmpttt.Denote£ of(2)by£ σ
amd
dellote£σ with σ=o,by£。and the correspondng semigroup by%(サ
).Letθし
)=:←
,μり,then we h訛Lemma l.FOr 9∈
σ(2(E))づιんοJ山 流αι£σ9=c g(£。一ψ
)(cg9),
磁θ ψ(μ
)=:((σ O,μ
3)̲(らμ2)2+(4②
σ,μ
2)+(Φりσ,μ
)― (らμ2)), σ(2)(″
,ν,Z)=σ(″ ,ν)σ
(ν,z),α
ηα Φ9σ(″)=σ(″,″ )α
ηαス(2)づsα
クしηθ%ιοr
cο
rrcΨ
οηttηθ ιο ιんC SC協な知η場(ι)づησ(E2)。P"J AccOrding to 131 ths theorem hOlds.
This implies the fO110wing
Theorem 4。 ■
ss%mc(c):σ
∈つ(4(2)),4(2)σ
∈σ(E2),α
ηα JCιつ(£σ)={9∈ σ(2(E)):♂9∈ つ(£。)}.
五ει{ル}bC α FJθ協づηナ協οι praccss cθttψθηごづり ιο(4o,つ(£o)).動
Cη ttc
づsts a scmむ御Ψ{γ(ι)}Cθ
rrcΨ
oηαづηクιο(£σ,つ(£σ))αηごづιんοJtt ιんαι γ(t)ψ(Jじ)==C gO・
)Eμlexp{ク(′
喝ヶ)‐―Qllヴ
,(″
ι3)as}9(ル)].P"J AccOrdng to Theorem 3.4 of pl£。gene酬ぃa Fen∝ semigroup.
乳 :器 ぎ潟
│ぷ説
l潟ぷ躙
;│け糖嚇観
'急百 ぷ
I訛鬱
ner乱∝
島 (ι
)9(μ)==Eμ
lexp{……
Jlι
ψ
(μ3)ご
S}9(ルι
t)]arld by Lemma l
К have fOr any λ>o(λ 一£σ) 1=c g(λ
̲fめ lcg,
so we obtain that
γ(ι)=C g%(サ)Cg.
TherefOre{γ
(サ)}iS a pOsitive semigroup.Becttse£σ is collselntive,{γ
(ι)}is a cOntrtttion semigroup. Q・
E.D。Ergodic properties ofneming― VIot pro∝sses with sdection 9 As the main result we halκ
Theorem 5。
4ss%鶴
c(C)αηα ttαι{%(t)}づ
S ιηοαづc αηα ιんαιルr Sοttι
′οsづ υc cθη
Sι
αηιs A√ ,λl αηα αstα
ιづοηarν αづsι
l陽b鶴
οtt H。‖ 名【
t)9‑(り,Io)1‖≦
yexp(一人
1サ)‖ ψ ll・
=ん
cη ι ん
c"づsお
α
sιαθ η
arνα づ stttbttι づ ο η
Ⅱ
sacん流α ιル r α ην
ε
>0流C 蒻st cθηstα
ηお ″1=″1(ε),δ =δ (ε
)>O Sαιり,加
ηクιんαι‖γ
(ι
)ψ ―(9,Ⅱ)1‖
≦ル名exp(―(λl―
ε)t)│lψ
‖ グ‖ψll<δ 。Fbr the proof the next lerlllnas on pertllrbation of operators are used.
Lenlina 2.Lct S bc a cθl叩
pα
ct,pacc αηα Πo bc α αづstttbttι
づοη οηS.■
ss鶴lmc
ιんαι Bづs a bttηαcα のcmιθr Oη L=σ(S)磁ιん鶴ηゆ 師 レηθttZ‖・‖and l̲3づsづηυ
cttblcづ
ηL.五cι P。
=(。,Πo)l
αηα y=P。 +3.ギびんas
αη cづθcη
υαJしclυづ流c勿
̀劉勉η
Ctづ
θη 9。,ι
んcη
りcん
αυc tんαι(a)9o=C(1‑3) 1l
α ηα〈
9o,Io)=C・(b)Lct υ
=(1‑3*) lΠo,ιん cη
υ
αρ α ηα
sα η
cづgcηψ
accげび
*Cοrrcψ θ η α づ η θ ι θ
cづgcηυ α :鶴 cl.
(C)野
磁η α α 流れο η
((1‑3) 21,I。 )≠ 0,Jθι
(9) Pl=((1‑3) 21,Ⅱ
。) 1(・
,(1‑3*) lⅡ o)(1‑3) 11,流
cη
び
Pl=Ply=Pl,
αηごPlづs a praFccιづο鶴・
(d)J‖ B‖
≦
:,流Cη
ιん
̀αssttηpιづοη げ
(c)。S Sα ιづ
Jttαηα
ttcη
(鵬づη 鶴α Jづ ιν
んοJごs
llび ―
Plll≦ 711 BII。
Praげ Because 9。 is alll eige]dhttLCtiOn,we haК
(9o,Π
o)1+39o=9o,so th乱
9o=(9o,Πo)(1‑3) 11
10 Seiich Π
WБ U
狙d((1‑3) 11,Πo)=1・
O is eigenvector ofび *cOrresporLding tO
(b)(a)implies th乱
(1‑3*) lΠIis arl eigenvector of D「 *,we haК
eigenvalue l. If
υi
(1,υ)Πo+B*υ =υ,
so that
υ
=(1,υ
)(1‑3*) lΠo。
TherefOre(b)hOldS.
(C)By(a)びPl=Pl holds and by(b)び*鍔 =鍔
hddS・
ObViously Pl Of(9)iS a prQ"ctiOno Let Bl=び一
Pl,then31=P。
―Pl+3.
(d)By l((1‑3) 21,H。
)‑11=│((1‑3) 231,Π
。)│≦ (1‑‖
BII) 2‖B‖, we halに,
陽 ―引 鋼Q.E.D。
Thё refOre the inequality holds.
Lemma 3。 助zαcr ttc assttηpιづθηげ
=ん
cο 鶴イυ
Cん
αυC ttαι「 亀(ι
)一
πKι)││≦C‖
ψ‖ι‑1.Praイ
πttt)=%(t)一 ル
(s)亀 (s)as
Jt%(ι ―S)1
ed by‖
%(ι
)││≦c‖
ψ‖ι Q.E.D。alld the inequality is obtain(
´
Praげ げ 動 協 J.Let P。 =(0,Π。
)1,then by the assumption ofthe theOrem we hatt fOr t。 =建14MII衝
【 ι
o)一 Po‖≦金
exp(一(λl―ε
)ιo)・
Then by LeIШma 3 we obtdn
lalt。
)一
%(ち)││≦金
eXp(―(λl―ε
)to)│
Ergodic properties ofneming‐
Vlot pЮ
cesses wim sdection llif‖ψll≦ δ≡ 寺10g(1+螢幽 讐 型 霊
)。
7(サ)iS COnservative,so by Theorem
4%(ι)haS tt eigenfunction c g correspondng to tt eigenvalue l.Put B=
πlto)一
Po・
Then‖B‖ ≦÷exp(―(λl―
ε)ιo)・ ByLema2(c)we haК
%(η
to)=Pl%+31π(η
>0)with some prttectiOrl Pl=(0,Π lルo SttiSfying 9。
=cc g Where c(≠
0)iS a constttt ttd Bl is a bollnded operator onσ (2(E))SttiStting
‖
3111≦ exp(―(λl― ε
)to)by Lemma 2(d)。
We C皿Obtain thtt for η t。 ≦ι
<(η +1)t0‖奮賓t)ψ―(9,Πl)9o‖
=‖
先(t一就。)(島 (η t。 )ψ
一(9,Ⅱl)9o)││≦
』4 exp(―(λl―ε)t)│19‖whereル亀=exp((λ
l―
ε十11ψ ll)to)。 SO the desired inequality is obtained bylettittg Aど1=埼‖c gll‖♂
│卜 Q・
E.D。4。 Conclusion and example
Assulne the mut乱
loll process is urlifordy ergodic,then by Theorem 2 arld TheoreⅡ1 5 the Fleming― Viot process with small selection is lⅡ
ufordyergodLc。
Let us now consider ttother formula with B=0証ld σ(″ ,ν)=(ん(″
)+
ん(ν))/2:
0・ 0=九 訳 ふ め く酬励幾 m
十薯
{ん ,ハ
+犠ん
,ハー 犠 フ 勁
0,ハ}:,α争 め 〉 粘
L貢ぶ
t「TWiβ 究 1亀
;Ψ't逸
践帯ゞ 星
fiきよ
M篤(C)beCOIIleS the colldition(D):ん
∈つ
(■),スん∈σ
(E)・ By Theorem 2狙 dTheoren1 5 we have
Theorem 6。
■
ss%mc(D)αη ご流α ι
{T(ι)}づ
S Cηο
ttc aηご流α ιル
r sοttθ
pοsづιづυc cοttstαη力9■√αηごλl αηα
α stαιづοηαη Jづ
stgttb鶴
Oη ι旬‖T(t)∫ ―(∫,均
)1‖
≦νexP(一人1サ)‖
∫‖・12 se五
ch口MnSU
動
cη
οηttc sc鶴
むratt γ(t)CοrrcΨ
θηαづηθ ιο(10)流θ 面StS a sιαιづοηarναづ
stttbttι
づθη Ⅱ sacん 流αιルr αην ε>Oιんc cttst cθηsι
αttts A4=』
イ1(ε),δ
=δ
(ε
)>O Sα 卵 ηJ ttαι‖γ(ι
),一
(9,Π)1‖
≦ル名exP(一(λl―
ε)サ)│lψ‖ ゲ‖ψll<δ.
Example l.Lct E=卜
1,1田dス =多 with theК
necting bollnda暉conditionつ
(ス)={ノ∈
(ア([‑1,ll),∫ ′
(‑1)‐
∫ ′
(1)=0},then the cOndition (A)of TheOrem 2 holds,
Praげ By Ю
l the trttsition density of{T(ι
)}iS in the formp(t,″,ν
)==:覇
フ霧亨Å lexp(一:〆ν―″+4η)2)=卜eXP(―
,甍:(ν+″■4η+2)2)]
=:十,Ic 翅
≒響範r2t sin工皇生
り
=」
主″Sin π(2η ‑1)ν+,Ic̲イ
:7r2t9。
sπη″COS πην・TherefoК
(A)お satiSied宙
th λl=子
.References
[lI Ethier,So N.arld G五
mths,R.C.The trttsition functiOn of a Fleming―
Viot process.■ ηη.Prab。 21(1993),1571‑1590。
レI Ethier,S.N.and KШ
tz,T.G.Mα
ttου PraccsscS,α しαttctcttzαιづοη αηJ.θοηυ
cη cη
cCO WileL New York(1986)。[31 Ethier,S,N.狙 d Kurtz,T.G.Fleming―ViOt prOcesses in populttion genetics.ЛИν J.a9ηtroJ ηJ o9 π.31(1993),345‐
386.
141 Ethier,S.N.alld KШ tz,T,Go Co■К
rgence to Fleming―
ViOt pЮcesses inthe weak atomic tOpdogy.膨
οcんαsι
づc Pracιsscs
ИppJ。.54(1994),1‑27.
bl Ethier,S.N.狙dK面 z,T,Go Couphng and ergodic theoreIIls for
Fleming̲ViOt pЮ cesses.■
ηη.Prab.26(1998),533‑561.
[61 Feller,W。,4η づηιroαttctづοη ιο probabづ
Jづ
ιν ιんcοη αηαづts Oppπ οηs.Vol.II.JOh WileyJ貶
SOns,Inc.,New York―
London―Sydney 1971
[71
同
Ergodic properties of Fleming-Viot processes with selection
Li,Z., Shiga, T., and Yao,L. A reversibility problem for Fleming-Viot processes. Electron'i,c Comnxun'ico,tions i,n Prob.4(1999) , 65-76
Shiga, T. A stochastic equation based on a Poisson system for a class
of measure-valued diffusion processes. J. Math. Kyoto Uni,u. 30(1990), 245-279
Tavar6 , S. Line-of-descent and genealogical processes, and their applica- tions in population genetics models. Thenret. Populati,on Bi,ol. 26(1984),
1 19-164.
回
14 seiich FrARU
自然淘汰を持つフレミント ウ オ過程のエルゴード的性質
板津
誠一
静岡大学理学部数学教室 ,
〒
422‐8529静岡市大谷
836突然変異の過程が一様 にエル ゴー ド的である条件の もとで,小さい自然淘汰 を持つ フレ ミ ングーヴ ィォ過程 が,一様 にエル ゴー ド的であることが示 され る.特に自然淘汰 を持 たない とき,突然変異の過程の分布 が指数的の速 さで定常分布 に収束す るときフレ ミングーヴ ィオ 過程の分布 も同 じ指数の速 さで定常分布 に収束す る。