Reflected Solutions of BSDEs Driven by G-Brownian Motion

Size: px
Start display at page:

Download "Reflected Solutions of BSDEs Driven by G-Brownian Motion"

Transcription

1 Refleced Soluion of BSDE Driven by G-Brownian Moion Hanwu Li Shige Peng Abdoulaye Soumana Hima Sepember 1, 17 Abrac In hi paper, we udy he refleced oluion of one-dimenional backward ochaic differenial equaion driven by G-Brownian moion (RGBSDE for hor). he reflecion keep he oluion above a given ochaic proce. In order o derive he uniquene of refleced G-BSDE, we apply a maringale condiion inead of he Skorohod condiion. Similar o he claical cae, we prove he exience by approximaion via penalizaion. Key word: G-expecaion, G-BSDE, refleced G-BSDE. MSC-claificaion: 6H1, 6H3 1 Inroducion El Karoui, Kapoudjian, Pardoux, Peng and Quenez [5] udied he problem of BSDE wih reflecion, which mean ha he oluion o a BSDE i required o be above a cerain given coninuou boundary proce, called he obacle. For hi purpoe, an addiional coninuou increaing proce hould be added in he equaion. Furhermore, hi addiional proce hould be choen in a minimal way o ha he Skorohod condiion i aified. An imporan obervaion i ha he oluion i he value funcion of an opimal opping problem. Due o he imporance in BSDE heory and in applicaion, he refleced problem ha araced a grea deal of aenion ince Many cholar ried o relax he condiion on he generaor and he obacle proce. Hamadene [8] and Lepelier and Xu [17] gave a generalized Skorohod condiion and proved he exience and uniquene when he obacle proce i no longer coninuou. Cvianic and Karaza [3] and Hamadene and Lepelier [9] udied he cae of wo reflecing obacle. hey alo eablihed he connecion beween hi problem and Dynkin game. Maoui [] and Kobylanki, Lepelier, Quenez and orre [16] exended he reul o he cae where he generaor i no a Lipchiz funcion. We hould poin ou ha he claical BSDE can only provide probabiliic inerpreaion for quailinear parial differenial equaion (PDE for hor). Beide, hi BSDE canno be applied o price pah-dependen coningen claim in he uncerain volailiy model (UVM for hor). Moivaed by hee fac, Peng [3, 4] yemeically inroduced a ime-conien fully nonlinear expecaion heory. One of he mo imporan cae i he G-expecaion heory (ee [7] and he reference School of Mahemaic, Shandong Univeriy, lihanwu@mail.du.edu.cn. School of Mahemaic and Qilu Iniue of Finance, Shandong Univeriy, peng@du.edu.cn. Li and Peng reearch wa parially uppored by he ian Yuan Projecion of he Naional Naure Science Foundaion of China (No and No ) and by he 111 Projec (No. B13). Li and Peng acknowledge graefully financial uppor by he German Reearch Foundaion (DFG) via CRC 183. Iniu de recherche mahémaique de Renne, Univerié de Renne 1 and Déparemen de mahémaique, Univerié de Maradi, oumanahima@yahoo.fr. A. Soumana Hima i graeful for parial financial uppor from he Lebegue Cener of Mahemaic ( Inveiemen d aveni Program) under gran ANR-11-LABX--1. 1

2 herein). In hi framework, a new ype of Brownian moion and he correponding ochaic calculu of Iô ype were conruced. I ha been widely ued o udy he problem of model uncerainy, nonlinear ochaic dynamical yem and fully nonlinear PDE. he backward ochaic differenial equaion driven by G-Brownian moion (i.e., G-BSDE) can be wrien in he following way Y = ξ + f(, Y, Z )d + g(, Y, Z )d B Z db (K K ). he oluion of hi equaion coni of a riple of procee (Y, Z, K). he exience and uniquene of he oluion are proved in [1]. In [11] he comparion heorem, Feymann-Kac formula and ome relaed opic aociaed wih hi kind of G-BSDE were eablihed. In hi paper, we udy he cae where he oluion of a G-BSDE i required o ay above a given ochaic proce, called he lower obacle. An increaing proce hould be added in hi equaion o puh he oluion upward, o ha i may remain above he obacle. According o he claical cae udied by [5], we may expec ha he oluion of refleced G-BSDE i a quadruple {(Y, Z, K, A ), } aifying (1) Y = ξ + f(, Y, Z )d + () (Y, Z, K) S α G (, ) and Y S, ; g(, Y, Z )d B Z db (K K ) + A A ; (3) {A } i coninuou and increaing, A = and (Y S )da =. he horcoming i ha he oluion of he above problem i no unique. hu, o ge he u- niquene of he refleced G-BSDE, we hould reformulae hi problem a he following. A riple of procee (Y, Z, A) i called a oluion of refleced G-BSDE if he following properie hold: (a) (Y, Z, A) S α G (, ) and Y S ; (b) Y = ξ + f(, Y, Z )d + g(, Y, Z )d B Z db + (A A ); (c) { (Y S )da } [, ] i a non-increaing G-maringale. Here, we denoe by SG α(, ) he collecion of proce (Y, Z, A) uch ha Y Sα G (, ), Z Hα G (, ), A i a coninuou nondecreaing proce wih A = and A SG α (, ). Noe ha we ue a maringale condiion (c) inead of he Skorohod condiion. Under ome appropriae aumpion, we can prove ha he oluion of he above refleced G-BSDE i unique. In proving he exience of hi problem, we hould ue he approximaion mehod via penalizaion. hi i a conrucive mehod in he ene ha he oluion of he refleced G-BSDE i proved o be he limi of a equence of penalized G-BSDE. Differen from he claical cae, he dominaed convergence heorem doe no hold under G-framework. Beide, any bounded equence in M p G (, ) i no longer weakly compac. he main difficuly in carrying ou hi conrucion i o prove he convergence propery in ome appropriae ene. I i worh poining ou ha he main idea i o apply he uniformly coninuou propery of he elemen in S p G (, ). Acually, he above equaion hold P -a.. for every probabiliy meaure P belong o a nondominaed cla of muually ingular meaure. herefore, he G-expecaion heory hare many imilariie wih econd order BSDE (BSDE for hor) developed by Cheridio, Soner, ouzi and Vicoir [1]. Maoui, Poamai and Zhou [1] howed he exience and uniquene of econd order refleced BSDE whoe oluion i (Y, Z, K P ) P P κ H aifying Y = ξ + ˆF (Y, Z )d Z db + (K P K P ), P -a..,

3 wih P P H (+,P ) EP Y S, K P k P = e inf P [K P k P ], P -a..,, P P κ H, where (y P, z p, k P ) denoe he unique oluion o he andard RBSDE wih daa (ξ, ˆF, S) under P. he main conribuion of our paper i ha he riple (Y, Z, A) i univerally defined wihin he G- framework uch ha he procee have rong regulariy propery. Due o hi propery, he oluion i ime-conien and he proce A can be aggregaed ino a univeral proce. Similar wih [5], when he refleced G-BSDE i formulaed under a Markovian framework, he oluion of hi problem provide a probabiliic repreenaion for he oluion of an obacle problem for nonlinear parabolic PDE. here ha been remendou inere in developing he obacle problem for parial differenial equaion ince i ha wide applicaion o mahemaical finance (ee [6]) and mahemaical phyic (ee [9]). he mehod in hi paper i called he Feynman-Kac formula which give a link beween probabiliy heory and PDE uing he language of vicoiy oluion. he oher approach i relaed o he variaional inequaliie and in hi cae he oluion belong o a Sobolev pace, ee in paricular [15] and [7]. We hould poin ou ha boh of he oluion udied by hee wo mehod are in weak ene. he re of paper i organized a follow. In Secion, we preen ome noaion and reul a preliminarie for he laer proof. he problem i formulaed in deail in Secion 3 and we ae ome eimae of he oluion from which we derive ome inegrabiliy properie of he oluion. In Secion 4, we eablih he approximaion mehod via penalizaion. We ae ome convergence properie of he oluion o he penalized G-BSDE. Our main reul are howed and proved in Secion 5. Furhermore, we prove a comparion heorem imilar o ha in [11], pecifically for nonrefleced G- BSDE. In Secion 6, we give he relaion beween refleced G-BSDE and he correponding obacle problem for fully nonlinear parabolic PDE. Finally, we ue he reul of he previou ecion o udy he pricing problem for American coningen claim under model uncerainy in Secion 7. In Appendix we inroduce he opional opping heorem under G-framework uing for pricing for American coningen claim. Preliminarie We recall ome baic noion and reul of G-expecaion, which are needed in he equel. relevan deail can be found in [1], [11], [5], [6], [7]. More.1 G-expecaion Definiion.1 Le Ω be a given e and le H be a vecor laice of real valued funcion defined on Ω, namely c H for each conan c and X H if X H. H i conidered a he pace of random variable. A ublinear expecaion Ê on H i a funcional Ê : H R aifying he following properie: for all X, Y H, we have (i) Monooniciy: If X Y, hen Ê[X] Ê[Y ]; (ii) Conan preerving: Ê[c] = c; (iii) Sub-addiiviy: Ê[X + Y ] Ê[X] + Ê[Y ]; (iv) Poiive homogeneiy: Ê[λX] = λê[x] for each λ. he riple (Ω, H, Ê) i called a ublinear expecaion pace. X H i called a random variable in (Ω, H, Ê). We ofen call Y = (Y 1,..., Y d ), Y i H a d-dimenional random vecor in (Ω, H, Ê). 3

4 Le Ω = C ([, ]; R d ), he pace of R d -valued coninuou funcion on [, ] wih ω =, be endowed wih he upremum norm, and B = (B i ) d i=1 be he canonical proce. For each >, denoe L ip (Ω ) := {ϕ(b 1,..., B n ) : n 1, 1,..., n [, ], ϕ C Lip (R d n )}. Denoe by S d he collecion of all d d ymmeric marice. For each given monoonic and ublinear funcion G : S d R, we can conruc a G-expecaion pace (Ω, L ip (Ω ), Ê, Ê). he canonical proce B i he d-dimenional G-Brownian moion under hi pace. In hi paper, we uppoe ha G i non-degenerae, i.e., here exi ome σ > uch ha G(A) G(B) 1 σ r[a B] for any A B. Le B be he d-dimenional G-Brownian moion. For each fixed a R d, {B a } := { a, B } i a 1-dimenional G a -Brownian moion, where G a : R R aifie G a (p) = G(aa )p + + G( aa )p. Le π N = { N,, N N }, N = 1,,, be a equence of pariion of [, ] uch ha µ(πn ) = max{ N i+1 N i : i =,, N 1}, he quadraic variaion proce of Ba i defined by N 1 B a = lim (B a µ(π N ) N j+1 j= B a ). N j For a, ā R d, we can define he muual variaion proce of B a and Bā by B a, Bā := 1 4 [ Ba+ā B a ā ]. Denoe by L p G (Ω ) he compleion of L ip (Ω ) under he norm ξ L p := (Ê[ ξ p ]) 1/p for p 1. For G. herefore i can be all [, ], Ê[ ] i a coninuou mapping on L ip (Ω ) w.r.. he norm L p G exended coninuouly o he compleion L p G (Ω ). Deni e al. [4] proved he following repreenaion heorem of G-expecaion on L 1 G (Ω ). heorem. ([4, 1]) here exi a weakly compac e P M 1 (Ω ), he e of all probabiliy meaure on (Ω, B(Ω )), uch ha P i called a e ha repreen Ê. Ê[ξ] = up E P [ξ] for all ξ L 1 G(Ω ). P P Le P be a weakly compac e ha repreen Ê. For hi P, we define capaciy c(a) := up P (A), A B(Ω ). P P Definiion.3 A e A B(Ω ) i polar if c(a) =. A propery hold quai-urely (q..) if i hold ouide a polar e. In he following, we do no diinguih wo random variable X and Y if X = Y q... For ξ L ip (Ω ), le E(ξ) = Ê[up [, ] Ê[ξ]], where Ê i he G-expecaion. For convenience, we call E G-evaluaion. For p 1 and ξ L ip (Ω ), define ξ p,e = [E( ξ p )] 1/p and denoe by L p E (Ω ) he compleion of L ip (Ω ) under p,e. he following eimae beween he wo norm L p and G p,e will be frequenly ued in hi paper. heorem.4 ([3]) For any α 1 and δ >, L α+δ G (Ω ) L α E (Ω ). More preciely, for any 1 < γ < β := (α + δ)/α, γ, we have ξ α α,e γ { ξ α L α+δ G where C β/γ = i=1 i β/γ,γ = γ/(γ 1) /γ C β/γ ξ (α+δ)/γ }, ξ L L α+δ ip (Ω ). G 4

5 . G-Iô calculu Definiion.5 Le MG (, ) be he collecion of procee in he following form: for a given pariion {,, N } = π of [, ], η (ω) = N 1 j= ξ j (ω)1 [j, j+1)(), where ξ i L ip (Ω i ), i =, 1,,, N 1. For each p 1 and η MG (, ) le η H p := G {Ê[( η d) p/ ]} 1/p, η M p := (Ê[ G η p d]) 1/p and denoe by H p G (, ), M p G (, ) he compleion of MG (, ) under he norm H p, G M p repecively. G For wo procee η MG (, ) and ξ M G 1 (, ), he G-Iô inegral ( η db) i and ( ξ d B i, B j ) are well defined, ee Li-Peng [19] and Peng [7]. Moreover, by Propoiion.1 in [19] and he claical Burkholder-Davi-Gundy inequaliy, he following propery hold. Propoiion.6 If η HG α(, ) wih α 1 and p (, α], hen we can ge up u [, ] u η db p L 1 G (Ω ) and σ p c p Ê [( η d) p/ ] Ê[ up u [, ] u η db p ] σ p C p Ê [( η d) p/ ]. Le SG (, ) = {h(, B 1,..., B n ) : 1,..., n [, ], h C b,lip (R n+1 )}. For p 1 and η SG (, ), e η S p = {Ê[up G [, ] η p ]} 1/p. Denoe by S p G (, ) he compleion of S G (, ) under he norm S p. We have he following coninuiy propery for any Y Sp G G (, ) wih p > 1. Lemma.7 ([18]) For Y S p G (, ) wih p > 1, we have, by eing Y := Y for >, F (Y ) := lim up(ê[ up ε up [, ] [,+ε] Y Y p ]) 1 p =. We now inroduce ome baic reul of G-BSDE. Conider he following ype of G-BSDE (here we ue Einein convenion) where Y = ξ + f(, Y, Z )d + g ij (, Y, Z )d B i, B j Z db (K K ), (.1) aifying he following properie: f(, ω, y, z), g ij (, ω, y, z) : [, ] Ω R R d R, (H1 ) here exi ome β > 1 uch ha for any y, z, f(,, y, z), g ij (,, y, z) M β G (, ), (H) here exi ome L > uch ha f(, y, z) f(, y, z ) + d g ij (, y, z) g ij (, y, z ) L( y y + z z ). i,j=1 For impliciy, we denoe by S α G (, ) he collecion of proce (Y, Z, K) uch ha Y Sα G (, ), Z HG α(, ; Rd ), K i a decreaing G-maringale wih K = and K L α G (Ω ). 5

6 heorem.8 ([1]) Aume ha ξ L β G (Ω ) and f, g ij aify (H1 ) and (H) for ome β > 1. hen for any 1 < α < β, equaion (.1) ha a unique oluion (Y, Z, K) S α G (, ). We alo have he comparion heorem for G-BSDE. heorem.9 ([11]) Le (Y l, Z, l K) l, l = 1,, be he oluion of he following G-BSDE: Y l = ξ l + f l (, Y l, Z)d l + gij(, l Y l, Z)d B l i, B j + V l V l ZdB l (K l K), l where {V l } (H), ξ l L β G (Ω ) wih β > 1. If ξ 1 ξ, f 1 f, gij 1 g ij increaing proce, hen Y 1 Y. are RCLL procee uch ha Ê[up [, ] V l β ] <, f l, g l ij aify (H1 ) and 1, for i, j = 1,, d, V V i an 3 Refleced G-BSDE wih a lower obacle and ome a priori eimae For impliciy, we conider he G-expecaion pace (Ω, L 1 G (Ω ), Ê) wih Ω = C ([, ], R) and σ = Ê[B 1] Ê[ B 1] = σ. Our reul and mehod ill hold for he cae d > 1. We are given he following daa: he generaor f and g, he obacle proce {S } [, ] and he erminal value ξ, where f and g are map f(, ω, y, z), g(, ω, y, z) : [, ] Ω R R. We will make he following aumpion: here exi ome β > uch ha (H1) for any y, z, f(,, y, z), g(,, y, z) M β G (, ); (H) f(, ω, y, z) f(, ω, y, z ) + g(, ω, y, z) g(, ω, y, z ) L( y y + z z ) for ome L > ; (H3) ξ L β G (Ω ) and ξ S, q..; (H4) here exi a conan c uch ha {S } [, ] S β G (, ) and S c, for each [, ]; (H4 ) {S } [, ] ha he following form S = S + b()d + l()d B + σ()db, where {b()} [, ], {l()} [, ] belong o M β G (, ) and {σ()} [, ] belong o H β G (, ). Le u now inroduce our refleced G-BSDE wih a lower obacle. A riple of procee (Y, Z, A) i called a oluion of refleced G-BSDE wih a lower obacle if for ome 1 < α β he following properie hold: (a) (Y, Z, A) S α G (, ) and Y S, ; (b) Y = ξ + f(, Y, Z )d + g(, Y, Z )d B Z db + (A A ); (c) { (Y S )da } [, ] i a non-increaing G-maringale. 6

7 Here we denoe by SG α(, ) he collecion of proce (Y, Z, A) uch ha Y Sα G (, ), Z Hα G (, ; R), A i a coninuou nondecreaing proce wih A = and A SG α (, ). For impliciy, we mainly conider he cae wih g and l. Similar reul ill hold for he cae g, l. Now we give a priori eimae for he oluion of he refleced G-BSDE wih a lower obacle. In he following, C will alway deignae a conan, which may vary from line o line. Propoiion 3.1 Le f aifie (H1) and (H). Aume Y = ξ + f(, Y, Z )d Z db + (A A ), where Y SG α(, ), Z Hα G (, ; R), A i a coninuou nondecreaing proce wih A = and A SG α (, ) for ome α > 1. hen here exi a conan C := C(α,, L, σ) > uch ha Ê [( Z d) α ] C{ Ê [ up Y α ] + (Ê[ up [, ] Ê [ A A α ] C{Ê[ up Y α ] + Ê[( [, ] [, ] Y α ]) 1/ (Ê[( f(,, ) d) α ]) 1/ }, (3.1) f(,, ) d) α ]}. (3.) Proof. he proof i imilar o ha of Propoiion 3.5 in [1]. So we omi i. Propoiion 3. For i = 1,, le ξ i L β G (Ω ), f i aify (H1) and (H) for ome β >. Aume Y i = ξ i + f i (, Y, Z )d Z i db + (A i A i ), where Y i SG α(, ), Zi HG α(, ), Ai i a coninuou nondecreaing proce wih A i = and A i SG α(, ) for ome 1 < α < β. Se Ŷ = Y 1 Y, Ẑ = Z 1 Z, Â = A 1 A. hen here exi a conan C := C(α,, L, σ) uch ha Ê[( Ẑ d) α ] Cα {(Ê[ up Ŷ α ]) 1/ [, ] + (Ê[( [(Ê[ up i=1 [, ] Y i α ]) 1/ f i (,, ) d) α ]) 1/ ] + Ê[ up Ŷ α ]}. [, ] Proof. he proof i imilar o ha of Propoiion 3.8 in [1]. So we omi i. Propoiion 3.3 For i = 1,, le ξ i L β G (Ω ) wih ξ i S i, where S i = S i + b i ()d + [, ] σ i ()db. Here {b i ()} M β G (, ), {σi ()} H β G (, ) for ome β >. Le f i aify (H1) and (H). Aume ha (Y i, Z i, A i ) SG α (, ) for ome 1 < α < β are he oluion of he refleced G-BSDE correponding o ξ i, f i and S i. Se Ỹ = (Y 1 S 1 ) (Y S ). hen here exi a conan C := C(α,, L, σ) uch ha Y i α CÊ[ ξ i α + up S i α + λ i, α d], Ỹ α CÊ[ ξ α + ( ˆλ α + ˆρ α + Ŝ α )d], where ξ = (ξ 1 S 1 ) (ξ S ), ˆλ = f 1 (, Y, Z ) f (, Y, Z ), ˆρ = b 1 () b () + σ 1 () σ (), Ŝ = S 1 S and = f i (,, ) + b i () + σ i (). λ i, 7

8 Proof. We will only how he econd inequaliy, ince he fir one can be proved in a imilar way. For any ε >, e ˆf = f 1 (, Y 1, Z 1 ) f (, Y, Z 1 ), ˆf = f 1 (, Y 1, Z 1 ) f 1 (, Y, Z ), Â = A 1 A, Z = (Z 1 σ 1 ()) (Z σ ()), ε α = ε(1 α/) + and Ȳ = Ỹ + ε α. Applying Iô formula o Ȳ α e r, where r > will be deermined laer, we ge Ȳ α/ e r + re r Ȳ α/ d + = (ε α + ξ ) α/ e r + α(1 α ) + α er Ȳ α/ 1 ( Z ) d B αe r Ȳ α/ 1 Ỹ ( ˆf + b 1 () b ())d + (ε α + ξ ) α/ e r + + α(1 α ) where M = Indeed, noe ha Conequenly, hen we obain αer Ȳ α/ 1 hu we can conclude ha e r Ȳ α/ (Ỹ) ( Z ) d B αe r Ȳ α/ 1 Ỹ dâ αe r Ȳ α 1 { ˆf 1 + b 1 () b () + ˆλ }d e r Ȳ α/ 1 ( Z ) d B (M M ), αe r Ȳ α/ 1 Ỹ Z db (3.3) (Ỹ Z db (Ỹ) + da 1 (Ỹ) da ). We claim ha {M } i a G-maringale. Ê[ Ỹ = Y 1 S 1 + S Y Y 1 S 1. (Ỹ) + (Y 1 S 1 ) + = Y 1 S 1. (Ỹ) + da 1 (Ỹ) + da 1 ] Ê[ (Y 1 S 1 )da 1. (Y 1 S 1 )da 1 ] =. I follow ha he proce {K 1 } [, ] = { (Ỹ) + da 1 } [, ] i a non-increaing G-maringale. Se {K } [, ] = { (Ỹ) da } [, ]. Boh {K 1 } and {K } are non-increaing G-maringale, o i { αer Ȳ α/ 1 (dk 1 + dk )}, which yield ha {M } [, ] i a G-maringale. From he aumpion of f 1, we derive ha αe r Ȳ α 1 ˆf 1 + b 1 () b () d αe r Ȳ α 1 {L( Ỹ + Z ) + (L 1)( Ŝ + ˆρ )}d (αl + αl σ (α 1) ) + (L 1) By Young inequaliy, we have 3(α 1) e r Ȳ α/ α(α 1) d + 4 αe r Ȳ α 1 { Ŝ + ˆρ }d. αe r Ȳ α 1 { ˆλ + Ŝ + ˆρ }d e r Ȳ α/ 1 ( Z ) d B e r Ȳ α/ d + e r { ˆλ α + ˆρ α + Ŝ α }d. 8 (3.4) (3.5)

9 By (3.3)-(3.5) and eing r = 3(L 1)(α 1) + αl + αl σ (α 1) + 1, we ge Ȳ α/ e r + (M M ) C{(ε α + ξ ) α/ e r + e r ( ˆλ α + ˆρ α + Ŝ α )d}. aking condiional expecaion on boh ide and hen by leing ε, we have he proof i complee. Ỹ α CÊ[ ξ α + ( ˆλ α + ˆρ α + Ŝ α )d]. Propoiion 3.4 Le (ξ, f, S) aify (H1)-(H4). Aume ha (Y, Z, A) SG α (, ), for ome α < β, i a oluion of he refleced G-BSDE wih daa (ξ, f, S). hen here exi a conan C := C(α,, L, σ, c) > uch ha Y α CÊ[1 + ξ α + f(,, ) α d]. Proof. For any r >, e Ỹ = Y c. Applying Iô formula o Ỹ α/ e r, noing ha S c and A i a nondecreaing proce, we have Ỹ α/ e r + = ξ c α e r + re r Ỹ α/ d + α e r Ỹ α/ 1 Z d B αe r Ỹ α/ 1 (Y c)f(, Y, Z )d + α(1 α ) αe r Ỹ α/ 1 (Y c)z db + ξ c α e r + αe r Ỹ α/ 1 (Y c)da e r Ỹ α/ (Y c) Z B αe r Ỹ α 1 f(, Y, Z ) d + α(1 α ) e r Ỹ α/ 1 Z B (M M ), where M = αe r Ỹ α/ 1 (Y c)z db αe r Ỹ α/ 1 (Y S )da. By condiion (c), M i a G-maringale. From he aumpion of f and by he Young inequaliy, we ge αe r Ỹ α 1 f(, Y, Z ) d Seing r = α + αl + αl σ (α 1) αe r Ỹ α 1 [ f(, c, ) + L Ỹ + L Z ]d (αl + αl σ (α 1) ) + α(α 1) 4 e r Ỹ α/ d + (α 1) e r Ỹ α/ 1 Z B + and by he above analyi, we have Ỹ α/ e r + M M ξ c α e r + aking condiional expecaion on boh ide yield ha Y c α CÊ[ ξ c α + e r f(, c, ) α d. f(, c, ) α d]. e r Ỹ α/ d e r f(, c, ) α d. Noing ha for p 1, we have a + b p p 1 ( a p + b p ). hen he proof i complee. (3.6) 9

10 Propoiion 3.5 Le (ξ 1, f 1, S 1 ) and (ξ, f, S ) be wo e of daa, each one aifying all he aumpion (H1)-(H4). Le (Y i, Z i, A i ) SG α (, ) be a oluion of he refleced G-BSDE wih daa (ξ i, f i, S i ), i = 1, repecively wih α < β. Se Ŷ = Y 1 Y, Ŝ = S 1 S, ˆξ = ξ 1 ξ. hen here exi a conan C := C(α,, L, σ, c) > uch ha Ŷ α C{Ê[ ˆξ α + where ˆλ = f 1 (, Y, Z ) f (, Y, Z ) and Ψ, = i=1 ˆλ α d] + (Ê[ up Ŝ α ]) 1 α 1 α Ψ α, }, [, ] Ê [ up Ê [1 + ξ i α + [, ] f i (r,, ) α dr]]. Proof. Se Ẑ = Z 1 Z, ˆf = f 1 (, Y 1, Z 1 ) f (, Y, Z ) and ˆf 1 = f 1 (, Y 1, Z 1 ) f 1 (, Y For any r >, by applying Iô formula o Ȳ α/ Ȳ α/ e r + re r Ȳ α/ α d + = ˆξ α e r + α(1 α ) + αe r Ȳ α/ 1 Ŷ ˆf d + ˆξ α e r + α(1 α ) + e r = ( Ŷ ) α/ e r, we have er Ȳ α/ 1 (Ẑ) d B e r Ȳ α/ (Ŷ) (Ẑ) d B αe r Ȳ α/ 1 Ŷ dâ e r Ȳ α/ 1 (Ẑ) d B + αe r Ȳ α 1 { ˆf 1 + ˆλ }d (M M ), αe r Ȳ α/ 1 Ŷ Ẑ db αe r Ȳ α/ 1 Ŝ dâ, Z ). where M = αer Ȳ α/ 1 Ŷ Ẑ db αer Ȳ α/ 1 (Ŷ Ŝ) da αer Ȳ α/ 1 (Ŷ Ŝ) + da 1. By a imilar analyi a he proof of Propoiion 3.3, we conclude ha {M } [, ] i a G-maringale. By Young inequaliy and he aumpion of f 1, imilar wih inequaliie (3.4) and (3.5), we have Se r = α + αl + αe r Ȳ α 1 { ˆf 1 α(α 1) + ˆλ }d 4 e r Ỹ α/ 1 Z B + + (α 1 + αl + αl σ (α 1) ) e r ˆλ α d e r Ỹ α/ d. αl σ (α 1). aking condiional expecaion on boh ide of (3.7), we obain Ŷ α C{Ê[ ˆξ α + By applying Hölder inequaliy, we ge Ê [ ˆλ α d] + Ê[ Ȳ α/ 1 Ŝ d(a 1 + A )]}. Ȳ α/ 1 Ŝ d(a 1 + A )] Ê[ up Ȳ α/ 1 Ŝ ( A 1 A 1 + A A )] [, ] (Ê[ up Ŝ α ]) 1 α ( Ê [ up [, ] [, ] Ȳ α/ ]) α α From Propoiion 3.1 and Propoiion 3.4, we finally ge he deired reul. ( Ê [ A i A i α ]) 1 α. i=1 (3.7) 1

11 Remark 3.6 If we require ha he oluion of a refleced G-BSDE i a quadruple {(Y, Z, K, A ), } aifying condiion (1)-(3) in he inroducion, he oluion i no unique. We can ee hi fac from he following example. 1 Le f 1, ξ = and S. I i eay o check ha (,,, ) and (,, σ σ (σ 1 B ), σ σ ( σ B )) are oluion of refleced G-BSDE wih daa (, 1, ) aifying all he condiion (1)-(3). 4 Penalized mehod and convergence properie In order o derive he exience of he oluion o he refleced G-BSDE wih a lower obacle, we hall apply he approximaion mehod via penalizaion. In hi ecion, we fir ae ome convergence properie of oluion o he penalized G-BSDE, which will be needed in he equel. For f and ξ aify (H1)-(H3), {S } [, ] aifie (H4) or (H4 ), we now conider he following family of G-BSDE parameerized by n = 1,,, Y n = ξ + f(, Y n, Z n )d + n (Y n S ) d Z n db (K n K n ). (4.1) Now le L n = n (Y n S ) d, hen (L n ) [, ] i a nondecreaing proce. We can rewrie refleced G-BSDE (4.1) a Y n = ξ + f(, Y n, Z n )d Z n db (K n K n ) + (L n L n ). (4.) We now eablih a priori eimae on he equence (Y n, Z n, K n, L n ) which are uniform in n. Lemma 4.1 here exi a conan C independen of n, uch ha for 1 < α < β, Ê[ up Y n α ] C, Ê[ K n α ] C, Ê[ L n α ] C, Ê[( Z n d) α ] C. [, ] Proof. For impliciy, fir we conider he cae S. he proof of he oher cae will be given in he remark. r, ε >, e Ỹ = (Y n ) + ε α, where ε α = ε(1 α/) +. Noe ha for any a R, a a. By applying Iô formula o Ỹ α/ e r yield ha Ỹ α/ e r + re r Ỹ α/ α d + er Ỹ α/ 1 (Z n ) d B = ( ξ + ε α ) α e r + α(1 α ) e r Ỹ α/ (Y n ) (Z n ) d B + αe r Ỹ α/ 1 Y n dl n + αe r Ỹ α/ 1 Y n f(, Y n, Z n )d αe r Ỹ α/ 1 (Y n Z n db + Y n dk n ) ( ξ + ε α ) α e r + α(1 α ) e r Ỹ α/ (Y n ) (Z n ) d B + where M = αe r Ỹ α/ 1/ f(, Y n, Z n ) d (M M ), αe r Ỹ α/ 1 (Y n Z n db +(Y n ) + dk n ) i a G-maringale. Similar wih inequaliy (3.6), 11

12 we have αe r Ỹ α 1 f(, Y n, Z n ) d e r f(,, ) α α(α 1) d + 4 Se r = α + αl + αl σ (α 1). We derive ha + (α 1 + αl + αl σ (α 1) ) Ỹ α/ e r + M M ( ξ + ε α ) α e r + e r Ỹ α/ 1 (Z n ) d B e r Ỹ α/ d. e r f(,, ) α d. aking condiional expecaion on boh ide and hen by leing ε, we obain Y n α CÊ[ ξ α + f(,, ) α d]. By heorem.4, for 1 < α < β, here exi a conan C independen of n uch ha Ê[up [, ] Y n α ] C. By Propoiion 3.1, we have Ê[( Z n d) α ] Cα {Ê[ up Y n α ] + (Ê[ up Y n α ]) 1 ( Ê[( [, ] [, ] Ê[ L n K n α ] C α {Ê[ up Y n α ] + Ê[( f(,, ) d) α ]}, [, ] f(,, ) d) α ]) 1 }, where he conan C α depend on α,, σ and L. hu we conclude ha here exi a conan C independen of n, uch ha for 1 < α < β, Since L n and Kn Ê[( Z n d) α ] C, Ê[ L n K n α ] C. are nonnegaive, i follow ha Ê[ K n α ] C, Ê[ L n α ] = n α Ê[( (Y n ) ) α ] C. Remark 4. If he obacle proce {S } [, ] aifie (H4), e Ỹ n ha = Y n c. I i imple o check Ỹ n = ξ c + f(, Ỹ n + c, Z n )d + n(ỹ n (S c)) d Z n db (K n K n ). By a imilar analyi a he proof of Lemma 4.1, we derive ha Ỹ n α CÊ[ ξ c α + f(, c, ) α d]. If S aifie (H4 ), for impliciy we uppoe ha l. Le Ỹ n we can rewrie (4.1) a he following: Ỹ n = ξ S + [f(, Ỹ n + S, Z n + σ()) + b()]d + n 1 = Y n S and Z n = Z n σ(), (Ỹ n ) d Z n db (K n K n ).

13 Uing he ame mehod, we ge Ỹ n α CÊ[ ξ S α + f(, S, σ()) + b() α d]. hu we conclude ha in he above wo cae, for 1 < α < β, here exi a conan C independen of n uch ha Ê[up [, ] Y n α ] C. By Propoiion 3.1, we have Ê[ K n α ] C, Ê[ L n α ] = n α Ê[( (Y n S ) d) α ] C, and Ê[( Z n d) α ] C. Lemma 4.1 implie ha (Y n S) in MG 1 (, ). he following lemma which correpond o Lemma 6.1 in [5] how ha hi convergence hold in SG α (, ), for 1 < α < β. I i of vial imporance o prove he convergence propery for (Y n ). Lemma 4.3 For ome 1 < α < β, we have lim Ê[ up (Y n S ) α ] =. n [, ] Proof. We now conider he following G-BSDE parameerized by n = 1,,, y n = ξ + f(, Y n, Z n )d + n(s y n )d z n db (k n k n ). By applying G-Iô formula o e n y n, we can ge y n = e n Ê [e n ξ + ne n S d + e n f(, Y n, Z n )d]. By he comparion heorem.9, we have for all n 1, Y n Y 1 and Y n S y n S = Ê[ S n + e n( ) f(, Y n, Z n )d], where S n = e n( ) (ξ S ) + ne n( ) (S S )d. I follow ha (Y n S ) (y n S ) Ê[ S n + e n( ) f(, Y n, Z n )d ]. Applying Hölder inequaliy yield ha e n( ) f(, Y n, Z n )d 1 ( f (, Y n, Z n )d) 1/ n By Lemma 4.1, for 1 < α < β, we have C ( up Y n n + [, ] f (,, ) + Z n d) 1/. Ê[ up [, ] e n( ) f(, Y n, Z n )d α ], a n. (4.3) 13

14 For ε >, i i raighforward o how ha S n = e n( ) (ξ S ) + ne n( ) (S S )d + +ε e n( ) ξ S + e nε up [+ε, ] For > δ >, from he above inequaliy we obain up S n e nδ up ξ S + e nε up [, δ] [, δ] +ε S S + up S S. [,+ε] up [, δ] [+ε, ] e nδ ( up S + ξ ) + e nε up S + up [, ] [, ] I i eay o check ha for each fixed ε, δ >, ne n( ) (S S )d S S + up up [, ] [,+ε] up [, δ] [,+ε] S S. S S Ê[ up S n β ] C{(e nβε nβδ + e )Ê[ up S β + ξ β ] + Ê[ up [, δ] [, ] CÊ[ up up [, ] [,+ε] S S β ], a n. up [, ] [,+ε] S S β ]} (4.4) For 1 < α < β and < δ <, we have Ê[ up (Y n S ) α ] Ê[ up [, ] Ê[ up {Ê[ S n + C{Ê[ [, δ] up [, δ] Ê [ up u [, δ] [, δ] (Y n S ) α ] + Ê[ up (Y n S ) α ] [ δ, ] e n( ) f(, Y n, Z n )d ]} α ] + Ê[ up [ δ, ] (Y 1 S ) α ] S u n α ]] + Ê[ up Ê [ up e n( ) f(, Y n, Z n )d α ]]} [, δ] u [, ] u + Ê[ up (Y 1 S ) α ] =: I + Ê[ up (Y 1 S ) α ]. [ δ, ] [ δ, ] (4.5) By Lemma.7, noing ha Y 1 S S α G (, ) and (Y 1 S ) =, we obain lim Ê[ up (Y 1 S ) α ] =. δ [ δ, ] By heorem.4 and combining (4.3), (4.4), we derive ha I C{Ê[ up up [, ] [,+ε] S S β ] + (Ê[ up up [, ] [,+ε] S S β ]) α/β }, a n. Now fir le n and hen le ε, δ in (4.5). By Lemma.7 again, he above analyi prove ha for 1 < α < β, lim Ê[ up (Y n S ) α ] =. n [, ] Now we how he convergence propery of equence (Y n ) n=1. Lemma 4.4 For ome β > α, we have lim Ê[ up Y n Y m α ] =. n,m [, ] 14

15 Proof. Wihou lo of generaliy, we may aume S in (4.1). For any r >, e Ȳ = Y n Y m, ˆf = f(, Y n, Z n ) f(, Y m, Z m ). By applying Iô formula o Ȳ α/ e r, we ge Ȳ α/ e r + re r Ȳ α/ α d + er Ȳ α/ 1 (Ẑ) d B where M = we have = α(1 α ) e r Ȳ α/ (Ŷ) (Ẑ) d B + + αe r Ȳ α/ 1 Ŷ ˆf d αe r Ȳ α/ 1 α(1 α ) e r Ȳ α/ (Ŷ) (Ẑ) d B + αer Ȳ α/ 1 αe r Ȳ α/ 1 Y n dl m αe r Ȳ α/ 1 Ŷ dˆl (ŶẐdB + Ŷd ˆK ) αe r Ȳ α 1 ˆf d αe r Ȳ α/ 1 Y m dl n (M M ), (ŶẐdB + (Ŷ) + dk m + (Ŷ) dk n ) i a G-maringale. Similar wih (3.4), αe r Ȳ α 1 ˆf d (αl + αl σ (α 1) ) Le r = 1 + αl + αl σ (α 1). By he above analyi, we have Ȳ α/ e r + (M M ) e r Ȳ α/ α(α 1) d + e r Ȳ α/ 1 (Ẑ) d B. 4 αe r Ȳ α/ 1 Y n dl m αe r Ȳ α/ 1 Y m dl n. hen aking condiional expecaion on boh ide of he above inequaliy, we conclude ha Oberve ha Ȳ α/ e r Ê[ αe r Ȳ α/ 1 Y n dl m αe r Ȳ α/ 1 Y m dl n ]. (4.6) Ê [ CÊ[ αe r Ȳ α/ 1 Y m dl n ] αe r Ê [ n (Y n ) α 1 (Y m ) d] + CÊ[ From (4.6) and aking expecaion on boh ide, we deduce ha Ȳ α/ 1 n(y n ) (Y m ) d] n (Y m ) α 1 (Y n ) d]. Ê[ up Y n [, ] Y m α ] CÊ[ up + Ê[ [, ] {Ê[ (n + m) (Y n ) α 1 (Y m ) d] (n + m) (Y m ) α 1 (Y n ) d]}]. (4.7) For α < β, here exi α, p, q, r, p, q > 1, uch ha 1 p + 1 q + 1 r = 1, 1 p + 1 q = 1, (α )α p < β, α q < β, α r < β, (α 1)α p < β and α q < β. Applying Lemma 4.1, Lemma 4.3 and he Hölder 15

16 inequaliy, here exi a conan C independen of m, n uch ha and Ê[( n (Y n ) α 1 (Y m ) d) α ] Ê[ up { (Y n ) (α )α (Y m ) α }( [, ] n(y n ) d) α ] (Ê[ up (Y n ) (α )α p ]) 1 p ( Ê[ up (Y m ) α q ]) 1 q ( Ê[( [, ] [, ] C(Ê[ up (Y m ) α q ]) 1 q, [, ] Ê[( m (Y n ) α 1 (Y m ) d) α ] Ê[ up (Y n ) (α 1)α ( [, ] m(y m ) d) α ] (Ê[ up (Y n ) (α 1)α p ]) 1 p (Ê[( m(y m ) d) α q ]) 1 q [, ] C(Ê[ up (Y n ) (α 1)α p ]) 1 p. [, ] hen by heorem.4 and Lemma 4.3, inequaliie (6.1)-(6.3) yield ha lim Ê[ up Y n Y m α ] =. n,m [, ] n(y n ) d) α r ]) 1 r (4.8) (4.9) 5 Exience and uniquene of refleced G-BSDE wih a lower obacle heorem 5.1 Suppoe ha ξ, f aify (H1)-(H3) and S aifie (H4) or (H4 ). hen he refleced G-BSDE wih daa (ξ, f, S) ha a unique oluion (Y, Z, A). Moreover, for any α < β we have Y SG α(, ), Z Hα G (, ) and A Sα G (, ). Proof. he uniquene of he oluion i a direc conequence of he a priori eimae in Propoiion 3., Propoiion 3.3 and Propoiion 3.5. o prove he exience, i uffice o prove he S cae. Recalling penalized G-BSDE (4.1), e Ŷ = Y n Y m, Ẑ = Z n Z m, ˆK = K n K m, ˆL = L n L m and ˆf = f(, Y n, Z n ) f(, Y m, Z m ). By Lemma 4.4, here exi Y SG α(, ) aifying lim n Ê[up [,] Y Y n α ] =. Applying Iô formula o Ŷ, we ge = Ŷ + L Ŷ ˆf d Ẑ d B Ŷd ˆK + [ Ŷ + Ŷ Ẑ ]d ŶdˆL Ŷd ˆK + 16 ŶẐdB ŶdˆL ŶẐdB.

17 Noe ha for each ε >, Chooing ε < σ, we have L Ŷ Ẑ d L /ε Ŷ d + ε Ẑ d. Ẑ d C( Ŷ d Ŷ d ˆK + Ŷ dˆl Ŷ Ẑ db ) C( up Ŷ + [, ] up [, ] Ŷ ( K n + K m + L n + L m ) Ŷ Ẑ db ). (5.1) By Propoiion.6, for any ε >, we obain Ê[( Ŷ Ẑ db ) α ] C Ê[( Ŷ d) α 4 ] C(Ê[ up Ŷ α ]) 1/ (Ê[( [, ] C Ê[ up Ŷ α ] + Cε Ê[( 4ε [, ] Ẑ d) α ]) 1/ Ẑ d) α ]. Applying Lemma 4.1 and he Hölder inequaliy, chooing ε mall enough, i follow from (5.1) ha I i raighforward o how ha Ê[( Z n Z m d) α ] C{ Ê[ up Ŷ α ] + (Ê[ up Ŷ α ]) 1/ }. [, ] [, ] lim Ê[( Z n Z m d) α ] =. n,m hen here exi a proce {Z } H α G (, ) uch ha Ê[( Z Z n d) α/ ] a n. Se A n = L n K n, i i eay o check ha (A n ) [, ] i a nondecreaing proce and A n A m = Ŷ Ŷ ˆf d + Ẑ db. By applying Propoiion.6 and he aumpion of f, i follow ha Ê[ up A n A m α ] CÊ[ up Ŷ α + ( ˆf d) α + up [, ] [, ] [, ] C{Ê[ up Ŷ α ] + Ê[( Ẑ d) α/ ]}. [, ] hen here exi a nondecreaing proce (A ) [, ] aifying ha lim Ê[ up A A n α ] =. n [, ] Ẑ db α ] 17

18 In he following i remain o prove ha Y, [, ] and { Y da } [, ] i a nonincreaing G-maringale. For he fir aemen, i can be deduced eaily from Lemma 4.3. Se K n := Y dk n. Since Y, and K n i a decreaing G-maringale, hen K n i a decreaing G-maringale. Noe ha where Ỹ m up [, ] Y da K n up [, ] + { up { [, ] Y da + (Y n Y )dk n + Y da n + Ỹ m d(a n A ) + Y n n(y n (Y n Y )da n ) d } (Y Ỹ m )d(a n A ) } + up Y Y n [ A n + K n ] + up (Y n ) L n [, ] [, ] =I + II + III + IV, = m 1 i= Y mi i [ m i,m i+1 ) () and m i = i m, i =, 1,, m. By imple analyi, we have m 1 Ê[I] Ê[ up Y ( A n A m m + An i+1 i+1 m i [, ] Ê[II] Ê[III] Ê[IV ] i= A m i )] m 1 (Ê[ up Y ]) 1/ {(Ê[ An A m m ]) 1/ + (Ê[ An i+1 i+1 m i [, ] i= (Ê[ up Y Ỹ m ]) 1/ {(Ê[ An ]) 1/ + (Ê[ A ]) 1/ }, [, ] (Ê[ up Y Y n ]) 1/ {(Ê[ An ]) 1/ + (Ê[ Kn ]) 1/ }, [, ] (Ê[ up (Y n ) ]) 1/ (Ê[ Ln ]) 1/. [, ] A m i ]) 1/ }, hen for each fixed m, fir le n end o infiniy, we conclude ha lim Ê[ up n [, ] Y da K n ] C(Ê[ up Y Ỹ m ]) 1/. [, ] By Lemma 3. in [1], ending m end o infiniy, we ge lim n Ê[up [, ] Y da K n ] =. I follow ha { Y da } i a non-increaing G-maringale. Furhermore, we have he following reul. heorem 5. Suppoe ha ξ, f and g aify (H1)-(H3), S aifie (H4) or (H4 ). hen he refleced G-BSDE wih daa (ξ, f, g, S) ha a unique oluion (Y, Z, A). Moreover, for any α < β we have Y SG α(, ), Z Hα G (, ) and A Sα G (, ). Proof. he proof i imilar o ha of heorem 5.1. We nex prove a comparion heorem, imilar o ha of [11] for non-refleced G-BSDE. he proof i baed on he approximaion mehod via penalizaion. 18

19 heorem 5.3 Le (ξ 1, f 1, g 1, S 1 ) and (ξ, f, g, S ) be wo e of daa. Suppoe S i aify (H4) or (H4 ), ξ i, f i and g i aify (H1)-(H3) for i = 1,. We furher aume in addiion he following: (i) ξ 1 ξ, q..; (ii) f 1 (, y, z) f (, y, z), g 1 (, y, z) g (, y, z), (y, z) R ; (iii) S 1 S,, q... Le (Y i, Z i, A i ) be a oluion of he refleced G-BSDE wih daa (ξ i, f i, g i, S i ), i = 1, repecively. hen Y 1 Y, q.. Proof. We fir conider he following G-BSDE parameerized by n = 1,,, y n = ξ 1 + f 1 (, y n, z n )d + g 1 (, y n, z n )d B + n(y n S 1 ) d z n db (K n K n ). By a imilar analyi a he proof of heorem 5.1, i follow ha lim n Ê[up [, ] Y 1 y n α ] =, where α < β. Noing ha (Y, Z, A ) i he oluion of he refleced G-BSDE wih daa (ξ, f, g, S ) and Y S,, we have Y = ξ + f (, Y, Z )d+ g (, Y, Z )d B + n(y S ) d Z db +(A A ). Applying heorem.9 yield Y y n, for all n N. Leing n, we conclude ha Y Y 1. Remark 5.4 Acually, A can be repreened a he um of wo nondecreaing procee A 1 and A uch ha and for any, Ê [ (Y S )da =, (S r Y r )da 1 r] = (S r Y r )da 1 r. (5.) Indeed, e A 1 = I {Y >S }da, A = I {Y =S }da. I i eay o check ha A = A 1 + A and A aifie he Skorohod condiion. We now how ha A 1 aifie (5.). Se K := (S Y )da. By heorem 5.1, K i a decreaing G-maringale and K L p G (Ω ) for ome 1 < p < β, [, ]. Chooe a equence of bounded, nonnegaive and Lipchiz coninuou funcion ϕ n (x) uch ha ϕ(x) I {x>}. Se K := ϕ n (Y S )d K = (S Y )ϕ n (Y S )da. Applying Lemma 3.4 in [1], We obain ha K i a decreaing G-maringale. Furhermore, we have ϕ n (Y S )d K (S Y )da 1 L 1 G (Ω ), 19

20 where L 1 G (Ω ) i defined in he Appendix. By he exended condiional G-expecaion defined in [13], we derive ha Ê [ (S r Y r )da 1 r] = lim Ê [ n = lim = n (S r Y r )ϕ n (Y r S r )da r ] (S r Y r )ϕ n (Y r S r )da r (S r Y r )da 1 r. 6 Relaion beween refleced G-BSDE and obacle problem for nonlinear parabolic PDE In hi ecion, we will give a probabiliic repreenaion for oluion of ome obacle problem for fully nonlinear parabolic PDE uing he refleced G-BSDE we have menioned in he above ecion. For hi purpoe, we need o pu he refleced G-BSDE in a Markovian framework. For each (, x) [, ] R d, le {X,x, } be he unique R d -valued oluion of he SDE driven by G-Brownian moion (here we ue Einein convenion): X,x = x + b(r, X,x r )dr + l ij (r, X,x r )d B i, B j r + We aume ha he daa (ξ, f, g, S) of he refleced G-BSDE ake he following form: ξ = φ(x,x ), f(, y, z) = f(, X,x, y, z), S = h(, X,x ), g ij (, y, z) = g ij (, X,x, y, z), σ i (r, X,x r )db i r. (6.1) where b : [, ] R d R d, l ij : [, ] R d R d, σ i : [, ] R d R d, φ : R d R, f, g ij : [, ] R d R R d R and h : [, ] R d R are deerminiic funcion and aify he following condiion: (A1) l ij = l ji and g ij = g ji for 1 i, j d; (A) b, l ij, σ i, f, g ij are coninuou in ; (A3) here exi a poiive ineger m and a conan L uch ha b(, x) b(, x ) + d l ij (, x) l ij (, x ) + i,j=1 d σ i (, x) σ i (, x ) L x x, φ(x) φ(x ) L(1 + x m + x m ) x x, d f(, x, y, z) f(, x, y, z ) + g ij (, x, y, z) g ij (, x, y, z ) i.j=1 i=1 L[(1 + x m + x m ) x x + y y + z z ]. (A4) h i uniformly coninuou w.r. (, x) and bounded from above, h(, x) Φ(x) for any x R d ; (A4 ) h belong o he pace C 1, Lip ([, ] Rd ) and h(, x) Φ(x) for any x R d, where C 1, Lip ([, ] R d ) i he pace of all funcion of cla C 1, ([, ] R d ) whoe parial derivaive of order le han or equal o and ielf are Lipchiz coninuou funcion wih repec o x.

21 We have he following eimae of G-SDE, which come from Chaper V of Peng [7]. Propoiion 6.1 ([7]) Le ξ, ξ L p G (Ω ; R d ) and p. hen we have, for each δ [, ], Ê [ up [,+δ] Ê [ up X,ξ [,+δ] where he conan C depend on L, G, p, d and. X,ξ p ] C ξ ξ p, Ê [ X,ξ +δ p ] C(1 + ξ p ), X,ξ ξ p ] C(1 + ξ p )δ p/, Proof. For he reader convenience, we give a brief proof here. I i eay o check ha {X,ξ } [, ], } [, ] M p G (, ; Rd ). By Propoiion.6, we have {X,ξ Ê [ up [,+δ] X,ξ CÊ[ ξ ξ p + C{ ξ ξ p + Ê[ C{ ξ ξ p + X,ξ p ] +δ +δ +δ X,ξ X,ξ Ê [ up Xr,ξ r [,] X,ξ p d + up [,+δ] +δ X,ξ p d] + Ê[( Xr,ξ p ]d}. (σ(r, Xr,ξ ) σ(r, Xr,ξ ))db r p ] X,ξ X,ξ d) p/ ]} By he Gronwall inequaliy, we ge he fir inequaliy. he oher can be proved imilarly. I follow from he previou reul ha for each (, x) [, ] R d, here exi a unique riple, Z,x, A,x ) [, ], which olve he following refleced G-BSDE: (Y,x (i) Y,x (ii) Y,x = φ(x,x ) + db r + A,x Z,x r f(r, X,x r h(, X,x ), ; (iii) {A,x non-increaing G-maringale., Y,x, Z,x )dr + r A,x, ; } i nondecreaing and coninuou, and { r g ij(r, X,x r (Y,x r, Y,x, Z,x )d B i, B j r r r h(r, X,x r ))da,x, } i a r We now conider he following obacle problem for a parabolic PDE. { min( u(, x) F (D xu, D x u, u, x, ), u(, x) h(, x)) =, (, x) (, ) R d, u(, x) = φ(x), x R d, (6.) where F (D xu, D x u, u, x, ) =G(H(D xu, D x u, u, x, )) + b(, x), D x u + f(, x, u, σ 1 (, x), D x u,, σ d (, x), D x u ), H(D xu, D x u, u, x, ) = D xuσ i (, x), σ j (, x) + D x u, l ij (, x) + g ij (, x, u, σ 1 (, x), D x u,, σ d (, x), D x u ). We need o conider oluion of he above PDE in he vicoiy ene. he be candidae o define he noion of vicoiy oluion i by uing he language of ub- and uper-je; (ee []). 1

22 Definiion 6. Le u C((, ) R d ) and (, x) (, ) R d. We denoe by P,+ u(, x) [he parabolic uperje of u a (, x)] he e of riple (p, q, X) R R d S d which are uch ha u(, y) u(, x) + p( ) + q, y x + 1 X(y x), y x + o( + y x ). Similarly, we define P, u(, x) [he parabolic ubje of u a (, x)] by P, u(, x) := P,+ ( u)(, x). hen we can give he definiion of he vicoiy oluion of he obacle problem (6.). Definiion 6.3 I can be aid ha u C([, ] R d ) i a vicoiy uboluion of (6.) if u(, x) φ(x), x R d, and a any poin (, x) (, ) R d, for any (p, q, X) P,+ u(, x), min(u(, x) h(, x), p F (X, q, u(, x), x, )). I can be aid ha u C([, ] R d ) i a vicoiy uperoluion of (6.) if u(, x) φ(x), x R d, and a any poin (, x) (, ) R d, for any (p, q, X) P, u(, x), min(u(, x) h(, x), p F (X, q, u(, x), x, )). u C([, ] R d ) i aid o be a vicoiy oluion of (6.) if i i boh a vicoiy uboluion and uperoluion. We now define u(, x) := Y,x, (, x) [, ] R d. (6.3) I i imporan o noe ha u(, x) i a deerminiic funcion. We claim ha u i a coninuou funcion. For impliciy, we only conider he cae g = in he nex hree lemma. he reul ill hold for he oher cae. Lemma 6.4 Le aumpion (A1)-(A3) and (A4 ) hold. For each [, ], x 1, x R d, we have u(, x 1 ) u(, x ) C(1 + x 1 m + x m ) x 1 x. Proof. From Propoiion 3.3, ince u(, x) i a deerminiic funcion, we have u(, x 1 ) u(, x ) C{Ê[ (φ(x,x1 + + f(, X,x1 + h(, X,x1 ) h(, X,x1, Y,x1, Z,x1 )) (φ(x,x ) h(, X,x )) ) f(, X,x, Y,x1, Z,x1 ) d b 1 () b () + l 1 ij() l ij() + σ 1 i () σ i () ) h(, X,x ) d] + h(, x 1 ) h(, x ) }, (6.4) where for k = 1,, b k () = h(, X,x k l k ij() = D x h(, X,x k ) + b(, X,x k ), D x h(, X,x k ), ), l ij (, X,x k σi k () = σ i (, X,x k ), D x h(, X,x k ). ) + 1 D xh(, X,x k )σ i (, X,x k ), σ j (, X,x k ),

23 Se ˆX = X,x1 X,x. By he aumpion (A3), (A4 ) and Propoiion 6.1, we have u(, x 1 ) u(, x ) C{Ê[(1 + he proof i complee. + k=1 Ê[(1 + X,x k m ) ˆX ] + k=1 Ê[(1 + k=1 X,x k m ) ˆX ]d X,x k ) ˆX ]d + Ê[ ˆX ]d + x 1 x } C(1 + x 1 m 4 m 4 + x ){(Ê[ up ˆX 4 ]) 1/ + x 1 x } [, ] C(1 + x 1 m 4 + x m 4 ) x 1 x. Lemma 6.5 Le aumpion (A1)-(A4) hold. For each [, ], x, x R d, we have u(, x 1 ) u(, x ) C{(1 + x 1 m + x m ) x 1 x + (1 + x 1 m+1 + x m+1 ) x 1 x }. Proof. From Propoiion 3.5 and Propoiion 6.1, by a imilar analyi wih he above lemma, we ge he deired reul. he following lemma ae ha u(, x) i coninuou wih repec o. Lemma 6.6 he funcion u(, x) i coninuou in. Proof. We only need o prove he cae where (A1)-(A3) and (A4 ) hold. he cae ha (A1)-(A4) hold can be proved in a imilar way. We define X,x := x, Y,x := Y,x, Z,x := and A,x := for. hen we define he obacle { S u,x h(, x) + u = b(, X,x )d + u l ij (, X,x )d B i, B j + u σ i (, X,x )db, i u (, ]; h(, x), u [, ], where b(, X,x lij (, X,x σ i (, X,x I i eay o check ha (Y,x ), f,x, S,x ) where (φ(x,x ) = h(, X,x ) + b(, X,x ), D x h(, X,x ), ) = D x h(, X,x ), l ij (, X,x ) + 1 D xh(, X,x )σ i (, X,x ), σ j (, X,x ), ) = σ i (, X,x ), D x h(, X,x )., Z,x, A,x ) [, ] i he oluion o he refleced G-BSDE wih daa f,x (, y, z) := I [, ] ()f(, X,x, y, z). Fix x R d, for 1, by Propoiion 3.3, we have where u( 1, x) u(, x) = Y 1,x Y,x C{Ê[ (φ(x1,x ) h(, X 1,x )) (φ(x,x ) h(, X,x )) + h( 1, x) h(, x) + ˆλ 1, () + ˆρ 1, () + h(, X 1,x ) h(, X,x ) d]}, ˆλ 1, () = I [1, ]()f(, X 1,x, Y,x, Z,x ) I [, ]()f(, X,x, Y,x, Z,x ), 3

24 and Se ˆX x = X 1,x ˆρ 1, () = I [1, ]() b(, X 1,x ) I [, ]() b(, X,x ) + I [1, ]() l ij (, X 1,x ) I [, ]() l ij (, X,x ) + I [1, ]() σ i (, X 1,x ) I [, ]() σ i (, X,x ). X,x. By Hölder inequaliy and aumpion (A3), (A4 ), we deduce ha u( 1, x) u(, x) C{Ê[(1 + X1,x m + X,x m ) ˆX x ] + h( 1, x) h(, x) + Ê[(1 + X 1,x m + X,x m ) ˆX x ]d Noe ha ˆX x,x 1,x = X + 1 Ê[1 + X 1,x (m+) 6 + Y,x ]d}. X,x, for [, ]. Applying Propoiion 6.1, i follow ha u( 1, x) u(, x) C{(1 + x (m+1) 3 ) h(, x) h( 1, x) }. he proof i complee. We will ue he approximaion of he refleced G-BSDE by penalizaion. For each (, x) [, ] R d, n N, le {(Y n,,x, Z n,,x, K n,,x ), } denoe he oluion of he G-BSDE: Y n,,x We define =φ(x,x ) + + n (Y n,,x r f(r, X,x r, Yr n,,x, Zr n,,x )dr + h(r, Xr,x )) dr g ij (r, X,x r u n (, x) := Y n,,x,, x R d., Y n,,x, Z n,,x )d B i, B j r r Zr n,,x db r (K n,,x K n,,x ),. By heorem 4.5 in [11], u n i he vicoiy oluion of he parabolic PDE { u n (, x) F n (Dxu n (, x), D x u n (, x), u n (, x), x, ) =, (, x) [, ] R d u n (, x) = φ(x), x R d, r (6.5) where F n (D xu, D x u, u, x, ) = F (D xu, D x u, u, x, ) + n(u h(, x)). heorem 6.7 he funcion u defined by (6.3) i he unique vicoiy oluion of he obacle problem (6.). Proof. From he reul of he previou ecion, for each (, x) [, ] R d, we obain u n (, x) u(, x), a n. By Propoiion 4., heorem 4.5 in [11] and Lemma 6.4-Lemma 6.6, u n and u are coninuou. hen by applying Dini heorem, he equence u n uniformly converge o u on compac e. We fir how ha u i a uboluion of (6.). For each fixed (, x) (, ) R d, le (p, q, X) P,+ u(, x). Wihou lo of generaliy, we may aume ha u(, x) > h(, x). By Lemma 6.1 in [], here exi equence n j, ( j, x j ) (, x), (p j, q j, X j ) P,+ u nj ( j, x j ), 4

25 uch ha (p j, q j, X j ) (p, q, X). Since u n i he vicoiy oluion o equaion (6.5), i follow ha for any j, p j F nj (X j, q j, u nj ( j, x j ), x j, j ). Noe ha u n i uniformly convergence on compac e, by he aumpion ha u(, x) > h(, x), we derive ha for j large enough u nj ( j, x j ) > h( j, x j ); herefore, ending j goe o infiniy in he above inequaliy yield: p F (X, q, u(, x), x, ). hen we conclude ha u i a uboluion of (6.). I remain o prove ha u i a uperoluion of (6.). For each fixed (, x) (, ) R d, and (p, q, X) P, u(, x). Noing ha {Y,x } [, ] i he oluion of refleced G-BSDE wih daa (ξ, f, g, S), where S = h(, X,x ), we have u(, x) = Y,x Applying Lemma 6.1 in [] again, here exi equence h(, x). n j, ( j, x j ) (, x), (p j, q j, X j ) P, u nj ( j, x j ), uch ha (p j, q j, X j ) (p, q, X). Since u n i he vicoiy oluion o equaion (6.5), we derive ha for any j, p j F nj (X j, q j, u nj ( j, x j ), x j, j ). herefore Sending j in he above inequaliy, we have p j F (X j, q j, u nj ( j, x j ), x j, j ). p F (X, q, u(, x), x, ), which implie ha u i a uperoluion of (6.). hu u i a vicoiy oluion of (6.). Analyi imilar o he proof of heorem 8.6 in [5] how ha here exi a mo one oluion of he obacle problem (6.) in he cla of coninuou funcion which grow a mo polynomially a infiniy. he proof i complee. 7 American opion under volailiy uncerainy Now le u conider he financial marke wih volailiy uncerainy. he marke model M i inroduced in [31] coniing of wo ae whoe dynamic are given by dγ = rγ d, γ = 1, ds = rs d + S db, S = x >, where r i a conan inere rae. he ae γ = (γ ) repreen a rikle bond. he ock price i decribed by a geomeric G-Brownian moion. Since he deviaion of he proce B from i mean i unknown, hi model how he ambiguiy under volailiy uncerainy. Definiion 7.1 ([31]) A cumulaive conumpion proce C = (C ) i a nonnegaive F -adaped proce wih value in L 1 G (Ω ), and wih nondecreaing, RCLL pah on (, ], and C =, C <, q... A porfolio proce π = (π ) i an F -adaped real valued proce wih value in L 1 G (Ω ). 5

26 Remark 7. Le (Ω, F, P ) be he claical probabiliy pace, where Ω = C ([, ], R) and F = B(Ω ). Furhermore, here exi a proce W = {W } which i a Brownian moion wih repec o P. hen he filraion in he above definiion i given by F = σ{w } N, where N denoe he collecion of all P -null ube. Definiion 7.3 ([31]) For a given iniial capial y, a porfolio proce π, and a cumulaive conumpion proce C, conider he wealh equaion dx = X (1 π ) dγ γ wih iniial wealh X = y. Or equivalenly γ 1 X = y + X π ds S = rx d + π X db dc γ 1 u dc u + dc γ 1 u X u π u db u,. If hi equaion ha a unique oluion X = (X ) := X y,π,c, i i called he wealh proce correponding o he riple (y, π, C). Definiion 7.4 ([31]) A porfolio/conumpion proce pair (π, C) i called admiible for an iniial capial y R if (i) he pair obey he condiion of Definiion 7.1 and 7.3; (ii) (π X y,π,c ) MG (, ); (iii) he oluion X y,π,c aifie X y,π,c L,, q.. where L i a nonnegaive random variable in L G (Ω ). We hen wrie (π, C) A(y). We denoe by, he e of all opping ime aking value in [, ], for any. hen he American coningen claim may be defined by he following: Definiion 7.5 ([14]) An American coningen claim i a financial inrumen coniing of (i) an expiraion dae (, ]; (ii) he elecion of an exercie ime τ, ; (iii) a payoff H τ a he exercie ime. We hould require ha he payoff proce {H } [, ] aify (H4) or (H4 ) in Secion 3. Since he financial marke under volailiy uncerainy i incomplee, i i naural o conider he uperhedging price for he American coningen claim. Definiion 7.6 Given an American coningen claim (, H) we define he uperhedging cla U := {y (π, C) A(y) : for any opping ime τ, X y,π,c τ he uperhedging price i defined a h up := inf{u y U}. H τ, q..}. 6

27 heorem 7.7 Given he financial marke M and an American coningen claim (, H), we have h up = Y, where Y = (Y ) i he oluion o he refleced G-BSDE wih parameer (H, f, H ) wih f(y) = ry. Proof. Le y U. By he definiion of U, here exi a pair (π, C) A(y) uch ha for any opping ime τ, Xτ y,π,c H τ. Applying Lemma 3.4, Lemma 4. and Lemma 4.3 in [19], we derive ha for any η MG (, ) τ Ê[ η db ] =. hen we obain y = Ê[y + τ Ê[y + τ = Ê[γ 1 τ γu 1 Xτ y,π,c γu 1 Xu y,π,c π u db u ] Xu y,π,c π u db u ] Ê[γ 1 τ H τ ]. τ γ 1 u dc u ] I follow ha h up up τ, Ê[γτ 1 H τ ]. Now we urn o prove he invere inequaliy. Conider he following refleced G-BSDE { Y = H ry d Z db + (A A ), Y H. By heorem 5.1, here exi a unique oluion (Y, Z, A) o he above equaion. Le C = A, π = Z Y. hen H τ Y τ = Xτ Y,π,C, which implie Y U. I follow ha h up Y. Applying Iô formula o Ỹ = γ 1 Y, we conclude ha Ỹ i a oluion o he refleced G-BSDE wih daa (γ 1 H,, γ 1 H). By he following propoiion, we finally ge he deired reul. Propoiion 7.8 Le (Y, Z, A) be a oluion of he refleced G-BSDE wih daa (ξ, f, S). hen we have Y = up Ê[ τ, τ Proof. Le τ,. Noe he fac ha hen we have f(, Y, Z )d + S τ I {τ< } + ξi {τ= } ]. τ Ê[ Z db ] =. τ Y = Ê[ f(, Y, Z )d + Y τ + A τ ] τ Ê[ f(, Y, Z )d + S τ I {τ< } + ξi {τ= } ]. We are now in a poiion o how he invere inequaliy. By he definiion of he oluion of he refleced G-BSDE, we may define K := (Y S )da, 7

28 hen K i a non-increaing G-maringale. Le D n = inf{ : Y S < 1 n }. By example A.4, D n i a -opping ime for n 1. I i eay o check ha D n D, where D = inf{ : Y S = }. Noing ha A i nondecreaing, by heorem A.5, i follow ha D n = Ê[K D n] = Ê[ (Y S )da ] 1 nê[ A Dn], which yield Ê[ A D n] =. By he coninuiy propery of A, we have Ê[ A D] =. hen i i eay o check ha D Y = Ê[ f(, Y, Z )d + S D I {D< } + ξi {D= } ]. Hence, he reul follow. A In hi ecion, we mainly inroduce he exended condiional G-expecaion and opional opping heorem under G-framework. More deail can be found in [13]. Le (Ω, L ip (Ω), Ê[ ]) be he G-expecaion pace and P be a weakly compac e ha repreen Ê. We e L (Ω) := {X : Ω [, ] and X i B(Ω)-meaurable}, L(Ω) := {X L (Ω) : E P [X] exi for each P P}. We exend G-expecaion Ê o L(Ω) and ill denoe i by Ê. For each X L(Ω), we define hen we give ome noaion Ê[X] = up E P [X]. P P L p (Ω) := {X L (Ω) : Ê[ X p ] < } for p 1, L 1 G (Ω) := {X L 1 (Ω) : X n L 1 G(Ω) uch ha X n X, q..}, L 1 G (Ω) := {X L1 (Ω) : X n L 1 G (Ω) uch ha X n X, q..}, L 1 G (Ω) := {X L1 (Ω) : X n L 1 G (Ω) uch ha Ê[ X n X ]}. Se Ω = {ω : ω Ω} for >. Similarly, we can define L (Ω ), L(Ω ), L p (Ω ), L 1 G (Ω ), L 1 G (Ω ) and L 1 G (Ω ) repecively. hen we can exend he condiional G-expecaion o pace L 1 G (Ω). Propoiion A.1 ([13]) For each X L 1 G (Ω), we have, for each P P, Ê [X] = e up P E Q [X F ], Q P(,P ) P -a.., where P(, P ) = {Q P : E Q [X] = E P [X], X L ip (Ω )}. 8

29 We now give he definiion of opping ime under G-expecaion framework. Definiion A. A random ime τ : Ω [, ) i called a -opping ime if I {τ } L 1 G (Ω ) for each. Definiion A.3 For a given -opping ime τ and ξ L 1 G (Ω), we define Êτ [ξ] := M τ, where M = Ê [ξ] for. We hen give an example of -opping ime. Example A.4 Le (X ) [, ] be an d-dimenional righ coninuou proce uch ha X L 1 G (Ω ) for [, ]. For each fixed cloed e F R d, we define hen τ i a -opping ime. τ = inf{ : X / F }. Now we inroduce he following opional opping heorem under G-framework. heorem A.5 ([13]) Suppoe ha G i non-degenerae. Le M = Ê[ξ] for, ξ L p G (Ω ) wih p > 1 and le σ, τ be wo -opping ime wih σ τ. hen M τ, M σ L 1 G (Ω) and M σ = Êσ[M τ ], q... Acknowledgmen he auhor are graeful o Yongheng Song and Falei Wang for heir help and many ueful dicuion. Reference [1] P. Cheridio, H. M. Soner, N. ouzi, and N. Vicoir, Second order backward ochaic differenial equaion and fully nonlinear parabolic PDE. Comm. Pure Appl. Mah., 6 (7), [] M. Crandall, H. Ihii and P. L. Lion, Uer guide o he vicoiy oluion of econd order parial differenial equaion. Bull. Amer. Mah. Soc., 7 (199), [3] J. Cvianic and I. Karaza, Backward ochaic differenial equaion wih reflecion and Dynkin game. Ann. Probab., 4(4) (1996), [4] L. Deni, M. Hu and S. Peng, Funcion pace and capaciy relaed o a ublinear expecaion: applicaion o G-Brownian moion pahe. Poenial Anal., 34 (11), [5] N. El Karoui, C. Kapoudjian, E. Pardoux, S. Peng and M.C. Quenez, Refleced oluion of backward SDE, and relaed obacle problem for PDE. he Annal of Probabiliy, 3() (1997), [6] N. El Karoui, E. Pardoux and M.C. Quenez, Refleced backward SDE and American opion. Numerical Mehod in Finance (Cambridge Univ. Pre), (1997), [7] D. Friedman, Variaional Principle and Free-boundary Problem. Second ediion. Rober E.Krieger Publihing Co., Inc., Malabar, FL, [8] S. Hamadene, Refleced BSDE wih diconinuou barrier and applicaion. Sochaic Sochaic Rep. 74(3-4) (),

Backward Stochastic Differential Equations and Applications in Finance

Backward Stochastic Differential Equations and Applications in Finance Backward Sochaic Differenial Equaion and Applicaion in Finance Ying Hu Augu 1, 213 1 Inroducion The aim of hi hor cae i o preen he baic heory of BSDE and o give ome applicaion in 2 differen domain: mahemaical

More information

Explicit form of global solution to stochastic logistic differential equation and related topics

Explicit form of global solution to stochastic logistic differential equation and related topics SAISICS, OPIMIZAION AND INFOMAION COMPUING Sa., Opim. Inf. Compu., Vol. 5, March 17, pp 58 64. Publihed online in Inernaional Academic Pre (www.iapre.org) Explici form of global oluion o ochaic logiic

More information

Stability in Distribution for Backward Uncertain Differential Equation

Stability in Distribution for Backward Uncertain Differential Equation Sabiliy in Diribuion for Backward Uncerain Differenial Equaion Yuhong Sheng 1, Dan A. Ralecu 2 1. College of Mahemaical and Syem Science, Xinjiang Univeriy, Urumqi 8346, China heng-yh12@mail.inghua.edu.cn

More information

Fractional Ornstein-Uhlenbeck Bridge

Fractional Ornstein-Uhlenbeck Bridge WDS'1 Proceeding of Conribued Paper, Par I, 21 26, 21. ISBN 978-8-7378-139-2 MATFYZPRESS Fracional Ornein-Uhlenbeck Bridge J. Janák Charle Univeriy, Faculy of Mahemaic and Phyic, Prague, Czech Republic.

More information

Mathematische Annalen

Mathematische Annalen Mah. Ann. 39, 33 339 (997) Mahemaiche Annalen c Springer-Verlag 997 Inegraion by par in loop pace Elon P. Hu Deparmen of Mahemaic, Norhweern Univeriy, Evanon, IL 628, USA (e-mail: elon@@mah.nwu.edu) Received:

More information

, the. L and the L. x x. max. i n. It is easy to show that these two norms satisfy the following relation: x x n x = (17.3) max

, the. L and the L. x x. max. i n. It is easy to show that these two norms satisfy the following relation: x x n x = (17.3) max ecure 8 7. Sabiliy Analyi For an n dimenional vecor R n, he and he vecor norm are defined a: = T = i n i (7.) I i eay o how ha hee wo norm aify he following relaion: n (7.) If a vecor i ime-dependen, hen

More information

Introduction to SLE Lecture Notes

Introduction to SLE Lecture Notes Inroducion o SLE Lecure Noe May 13, 16 - The goal of hi ecion i o find a ufficien condiion of λ for he hull K o be generaed by a imple cure. I urn ou if λ 1 < 4 hen K i generaed by a imple curve. We will

More information

arxiv: v1 [math.pr] 2 Jan 2015

arxiv: v1 [math.pr] 2 Jan 2015 Mean-field backward ochaic differenial equaion on Markov chain arxiv:151.955v1 [mah.pr] 2 Jan 215 Wen Lu 1 Yong Ren 2 1. School of Mahemaic and Informaional Science, Yanai Univeriy, Yanai 2645, China 2.

More information

FULLY COUPLED FBSDE WITH BROWNIAN MOTION AND POISSON PROCESS IN STOPPING TIME DURATION

FULLY COUPLED FBSDE WITH BROWNIAN MOTION AND POISSON PROCESS IN STOPPING TIME DURATION J. Au. Mah. Soc. 74 (23), 249 266 FULLY COUPLED FBSDE WITH BROWNIAN MOTION AND POISSON PROCESS IN STOPPING TIME DURATION HEN WU (Received 7 Ocober 2; revied 18 January 22) Communicaed by V. Sefanov Abrac

More information

6.8 Laplace Transform: General Formulas

6.8 Laplace Transform: General Formulas 48 HAP. 6 Laplace Tranform 6.8 Laplace Tranform: General Formula Formula Name, ommen Sec. F() l{ f ()} e f () d f () l {F()} Definiion of Tranform Invere Tranform 6. l{af () bg()} al{f ()} bl{g()} Lineariy

More information

ESTIMATES FOR THE DERIVATIVE OF DIFFUSION SEMIGROUPS

ESTIMATES FOR THE DERIVATIVE OF DIFFUSION SEMIGROUPS Elec. Comm. in Probab. 3 (998) 65 74 ELECTRONIC COMMUNICATIONS in PROBABILITY ESTIMATES FOR THE DERIVATIVE OF DIFFUSION SEMIGROUPS L.A. RINCON Deparmen of Mahemaic Univeriy of Wale Swanea Singleon Par

More information

ANSWERS TO ODD NUMBERED EXERCISES IN CHAPTER

ANSWERS TO ODD NUMBERED EXERCISES IN CHAPTER John Riley 6 December 200 NWER TO ODD NUMBERED EXERCIE IN CHPTER 7 ecion 7 Exercie 7-: m m uppoe ˆ, m=,, M (a For M = 2, i i eay o how ha I implie I From I, for any probabiliy vecor ( p, p 2, 2 2 ˆ ( p,

More information

NECESSARY AND SUFFICIENT CONDITIONS FOR LATENT SEPARABILITY

NECESSARY AND SUFFICIENT CONDITIONS FOR LATENT SEPARABILITY NECESSARY AND SUFFICIENT CONDITIONS FOR LATENT SEPARABILITY Ian Crawford THE INSTITUTE FOR FISCAL STUDIES DEPARTMENT OF ECONOMICS, UCL cemmap working paper CWP02/04 Neceary and Sufficien Condiion for Laen

More information

Generalized Snell envelope and BSDE With Two general Reflecting Barriers

Generalized Snell envelope and BSDE With Two general Reflecting Barriers 1/22 Generalized Snell envelope and BSDE Wih Two general Reflecing Barriers EL HASSAN ESSAKY Cadi ayyad Universiy Poly-disciplinary Faculy Safi Work in progress wih : M. Hassani and Y. Ouknine Iasi, July

More information

Multidimensional Markovian FBSDEs with superquadratic

Multidimensional Markovian FBSDEs with superquadratic Mulidimenional Markovian FBSDE wih uperquadraic growh Michael Kupper a,1, Peng Luo b,, Ludovic angpi c,3 December 14, 17 ABSRAC We give local and global exience and uniquene reul for mulidimenional coupled

More information

An Introduction to Malliavin calculus and its applications

An Introduction to Malliavin calculus and its applications An Inroducion o Malliavin calculus and is applicaions Lecure 5: Smoohness of he densiy and Hörmander s heorem David Nualar Deparmen of Mahemaics Kansas Universiy Universiy of Wyoming Summer School 214

More information

BSDE Approach to Non-Zero-Sum Stochastic Differential Games of Control and Stopping

BSDE Approach to Non-Zero-Sum Stochastic Differential Games of Control and Stopping BSDE Approach o Non-Zero-Sum Sochaic Differenial Game of Conrol and Sopping Ioanni Karaza INTECH Invemen Managemen One Palmer Square, Suie 441 Princeon, NJ 8542 ik@enhanced.com Qinghua Li Saiic Deparmen,

More information

An Introduction to Backward Stochastic Differential Equations (BSDEs) PIMS Summer School 2016 in Mathematical Finance.

An Introduction to Backward Stochastic Differential Equations (BSDEs) PIMS Summer School 2016 in Mathematical Finance. 1 An Inroducion o Backward Sochasic Differenial Equaions (BSDEs) PIMS Summer School 2016 in Mahemaical Finance June 25, 2016 Chrisoph Frei cfrei@ualbera.ca This inroducion is based on Touzi [14], Bouchard

More information

FUZZY n-inner PRODUCT SPACE

FUZZY n-inner PRODUCT SPACE Bull. Korean Mah. Soc. 43 (2007), No. 3, pp. 447 459 FUZZY n-inner PRODUCT SPACE Srinivaan Vijayabalaji and Naean Thillaigovindan Reprined from he Bullein of he Korean Mahemaical Sociey Vol. 43, No. 3,

More information

Dual Representation as Stochastic Differential Games of Backward Stochastic Differential Equations and Dynamic Evaluations

Dual Representation as Stochastic Differential Games of Backward Stochastic Differential Equations and Dynamic Evaluations arxiv:mah/0602323v1 [mah.pr] 15 Feb 2006 Dual Represenaion as Sochasic Differenial Games of Backward Sochasic Differenial Equaions and Dynamic Evaluaions Shanjian Tang Absrac In his Noe, assuming ha he

More information

Laplace Transform. Inverse Laplace Transform. e st f(t)dt. (2)

Laplace Transform. Inverse Laplace Transform. e st f(t)dt. (2) Laplace Tranform Maoud Malek The Laplace ranform i an inegral ranform named in honor of mahemaician and aronomer Pierre-Simon Laplace, who ued he ranform in hi work on probabiliy heory. I i a powerful

More information

Rough Paths and its Applications in Machine Learning

Rough Paths and its Applications in Machine Learning Pah ignaure Machine learning applicaion Rough Pah and i Applicaion in Machine Learning July 20, 2017 Rough Pah and i Applicaion in Machine Learning Pah ignaure Machine learning applicaion Hiory and moivaion

More information

On the Exponential Operator Functions on Time Scales

On the Exponential Operator Functions on Time Scales dvance in Dynamical Syem pplicaion ISSN 973-5321, Volume 7, Number 1, pp. 57 8 (212) hp://campu.m.edu/ada On he Exponenial Operaor Funcion on Time Scale laa E. Hamza Cairo Univeriy Deparmen of Mahemaic

More information

arxiv: v1 [math.pr] 19 Nov 2018

arxiv: v1 [math.pr] 19 Nov 2018 Muli-dimenional BSDE wih diagonal generaor driven by G-Brownian moion arxiv:1811.7773v1 [mah.pr] 19 Nov 218 Guomin Liu November 2, 218 Abrac In hi paper, we udy he well-poedne of muli-dimenional backward

More information

EECE 301 Signals & Systems Prof. Mark Fowler

EECE 301 Signals & Systems Prof. Mark Fowler EECE 31 Signal & Syem Prof. Mark Fowler Noe Se #27 C-T Syem: Laplace Tranform Power Tool for yem analyi Reading Aignmen: Secion 6.1 6.3 of Kamen and Heck 1/18 Coure Flow Diagram The arrow here how concepual

More information

FLAT CYCLOTOMIC POLYNOMIALS OF ORDER FOUR AND HIGHER

FLAT CYCLOTOMIC POLYNOMIALS OF ORDER FOUR AND HIGHER #A30 INTEGERS 10 (010), 357-363 FLAT CYCLOTOMIC POLYNOMIALS OF ORDER FOUR AND HIGHER Nahan Kaplan Deparmen of Mahemaic, Harvard Univeriy, Cambridge, MA nkaplan@mah.harvard.edu Received: 7/15/09, Revied:

More information

Randomized Perfect Bipartite Matching

Randomized Perfect Bipartite Matching Inenive Algorihm Lecure 24 Randomized Perfec Biparie Maching Lecurer: Daniel A. Spielman April 9, 208 24. Inroducion We explain a randomized algorihm by Ahih Goel, Michael Kapralov and Sanjeev Khanna for

More information

Research Article Existence and Uniqueness of Solutions for a Class of Nonlinear Stochastic Differential Equations

Research Article Existence and Uniqueness of Solutions for a Class of Nonlinear Stochastic Differential Equations Hindawi Publihing Corporaion Abrac and Applied Analyi Volume 03, Aricle ID 56809, 7 page hp://dx.doi.org/0.55/03/56809 Reearch Aricle Exience and Uniquene of Soluion for a Cla of Nonlinear Sochaic Differenial

More information

Utility maximization in incomplete markets

Utility maximization in incomplete markets Uiliy maximizaion in incomplee markes Marcel Ladkau 27.1.29 Conens 1 Inroducion and general seings 2 1.1 Marke model....................................... 2 1.2 Trading sraegy.....................................

More information

Research Article On Double Summability of Double Conjugate Fourier Series

Research Article On Double Summability of Double Conjugate Fourier Series Inernaional Journal of Mahemaic and Mahemaical Science Volume 22, Aricle ID 4592, 5 page doi:.55/22/4592 Reearch Aricle On Double Summabiliy of Double Conjugae Fourier Serie H. K. Nigam and Kuum Sharma

More information

arxiv: v5 [math.pr] 6 Oct 2015

arxiv: v5 [math.pr] 6 Oct 2015 ON THE EXISTENCE AND UNIQUENESS OF SOLUTIONS TO STOCHASTIC DIFFERENTIAL EQUATIONS DRIVEN BY G-BROWNIAN MOTION WITH INTEGRAL-LIPSCHITZ COEFFICIENTS XUEPENG BAI AND YIQING LIN arxiv:12.146v5 mah.pr] 6 Oc

More information

Approximation for Option Prices under Uncertain Volatility

Approximation for Option Prices under Uncertain Volatility Approximaion for Opion Price under Uncerain Volailiy Jean-Pierre Fouque Bin Ren February, 3 Abrac In hi paper, we udy he aympoic behavior of he wor cae cenario opion price a he volailiy inerval in an uncerain

More information

Generalized Orlicz Spaces and Wasserstein Distances for Convex-Concave Scale Functions

Generalized Orlicz Spaces and Wasserstein Distances for Convex-Concave Scale Functions Generalized Orlicz Space and Waerein Diance for Convex-Concave Scale Funcion Karl-Theodor Surm Abrac Given a ricly increaing, coninuou funcion ϑ : R + R +, baed on he co funcional ϑ (d(x, y dq(x, y, we

More information

Backward stochastic differential equations and integral partial differential equations

Backward stochastic differential equations and integral partial differential equations Backward ochaic differenial equaion and inegral parial differenial equaion Guy BARLS, Rainer BUCKDAHN 1 and ienne PARDOUX 2 June 30, 2009 Abrac We conider a backward ochaic differenial equaion, whoe he

More information

arxiv: v2 [math.pr] 18 Mar 2019

arxiv: v2 [math.pr] 18 Mar 2019 Probabiliic approach o ingular perurbaion of vicoiy oluion o nonlinear parabolic PDE arxiv:193.5549v2 mah.pr 18 Mar 219 Minghang Hu Falei Wang Abrac In hi paper, we prove a convergence heorem for ingular

More information

Flow networks. Flow Networks. A flow on a network. Flow networks. The maximum-flow problem. Introduction to Algorithms, Lecture 22 December 5, 2001

Flow networks. Flow Networks. A flow on a network. Flow networks. The maximum-flow problem. Introduction to Algorithms, Lecture 22 December 5, 2001 CS 545 Flow Nework lon Efra Slide courey of Charle Leieron wih mall change by Carola Wenk Flow nework Definiion. flow nework i a direced graph G = (V, E) wih wo diinguihed verice: a ource and a ink. Each

More information

Coupling Homogenization and Large Deviations. Principle in a Parabolic PDE

Coupling Homogenization and Large Deviations. Principle in a Parabolic PDE Applied Mahemaical Science, Vol. 9, 215, no. 41, 219-23 HIKARI Ld, www.m-hikari.com hp://dx.doi.org/1.12988/am.215.5169 Coupling Homogenizaion and Large Deviaion Principle in a Parabolic PDE Alioune Coulibaly,

More information

1 Motivation and Basic Definitions

1 Motivation and Basic Definitions CSCE : Deign and Analyi of Algorihm Noe on Max Flow Fall 20 (Baed on he preenaion in Chaper 26 of Inroducion o Algorihm, 3rd Ed. by Cormen, Leieron, Rive and Sein.) Moivaion and Baic Definiion Conider

More information

Introduction to Congestion Games

Introduction to Congestion Games Algorihmic Game Theory, Summer 2017 Inroducion o Congeion Game Lecure 1 (5 page) Inrucor: Thoma Keelheim In hi lecure, we ge o know congeion game, which will be our running example for many concep in game

More information

Local Strict Comparison Theorem and Converse Comparison Theorems for Reflected Backward Stochastic Differential Equations

Local Strict Comparison Theorem and Converse Comparison Theorems for Reflected Backward Stochastic Differential Equations arxiv:mah/07002v [mah.pr] 3 Dec 2006 Local Sric Comparison Theorem and Converse Comparison Theorems for Refleced Backward Sochasic Differenial Equaions Juan Li and Shanjian Tang Absrac A local sric comparison

More information

ON FRACTIONAL ORNSTEIN-UHLENBECK PROCESSES

ON FRACTIONAL ORNSTEIN-UHLENBECK PROCESSES Communicaion on Sochaic Analyi Vol. 5, No. 1 211 121-133 Serial Publicaion www.erialpublicaion.com ON FRACTIONAL ORNSTEIN-UHLENBECK PROCESSES TERHI KAARAKKA AND PAAVO SALMINEN Abrac. In hi paper we udy

More information

EE Control Systems LECTURE 2

EE Control Systems LECTURE 2 Copyrigh F.L. Lewi 999 All righ reerved EE 434 - Conrol Syem LECTURE REVIEW OF LAPLACE TRANSFORM LAPLACE TRANSFORM The Laplace ranform i very ueful in analyi and deign for yem ha are linear and ime-invarian

More information

Algorithmic Discrete Mathematics 6. Exercise Sheet

Algorithmic Discrete Mathematics 6. Exercise Sheet Algorihmic Dicree Mahemaic. Exercie Shee Deparmen of Mahemaic SS 0 PD Dr. Ulf Lorenz 7. and 8. Juni 0 Dipl.-Mah. David Meffer Verion of June, 0 Groupwork Exercie G (Heap-Sor) Ue Heap-Sor wih a min-heap

More information

MALLIAVIN CALCULUS FOR BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS AND APPLICATION TO NUMERICAL SOLUTIONS

MALLIAVIN CALCULUS FOR BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS AND APPLICATION TO NUMERICAL SOLUTIONS The Annal of Applied Probabiliy 211, Vol. 21, No. 6, 2379 2423 DOI: 1.1214/11-AAP762 Iniue of Mahemaical Saiic, 211 MALLIAVIN CALCULUS FOR BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS AND APPLICATION TO

More information

FIXED POINTS AND STABILITY IN NEUTRAL DIFFERENTIAL EQUATIONS WITH VARIABLE DELAYS

FIXED POINTS AND STABILITY IN NEUTRAL DIFFERENTIAL EQUATIONS WITH VARIABLE DELAYS PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 136, Number 3, March 28, Page 99 918 S 2-9939(7)989-2 Aricle elecronically publihed on November 3, 27 FIXED POINTS AND STABILITY IN NEUTRAL DIFFERENTIAL

More information

T-Rough Fuzzy Subgroups of Groups

T-Rough Fuzzy Subgroups of Groups Journal of mahemaic and compuer cience 12 (2014), 186-195 T-Rough Fuzzy Subgroup of Group Ehagh Hoeinpour Deparmen of Mahemaic, Sari Branch, Ilamic Azad Univeriy, Sari, Iran. hoienpor_a51@yahoo.com Aricle

More information

An introduction to the (local) martingale problem

An introduction to the (local) martingale problem An inroducion o he (local) maringale problem Chri Janjigian Ocober 14, 214 Abrac Thee are my preenaion noe for a alk in he Univeriy of Wiconin - Madion graduae probabiliy eminar. Thee noe are primarily

More information

6. Stochastic calculus with jump processes

6. Stochastic calculus with jump processes A) Trading sraegies (1/3) Marke wih d asses S = (S 1,, S d ) A rading sraegy can be modelled wih a vecor φ describing he quaniies invesed in each asse a each insan : φ = (φ 1,, φ d ) The value a of a porfolio

More information

Uniqueness of solutions to quadratic BSDEs. BSDEs with convex generators and unbounded terminal conditions

Uniqueness of solutions to quadratic BSDEs. BSDEs with convex generators and unbounded terminal conditions Recalls and basic resuls on BSDEs Uniqueness resul Links wih PDEs On he uniqueness of soluions o quadraic BSDEs wih convex generaors and unbounded erminal condiions IRMAR, Universié Rennes 1 Châeau de

More information

About the Pricing Equation in Finance

About the Pricing Equation in Finance Abou he Pricing Equaion in Finance Séphane Crépey Déparemen de Mahémaique Univerié d Évry Val d Eonne 91025 Évry Cedex, France Thi verion: July 10, 2007 1 Inroducion In [16], we conruced a raher generic

More information

Fractional Brownian Bridge Measures and Their Integration by Parts Formula

Fractional Brownian Bridge Measures and Their Integration by Parts Formula Journal of Mahemaical Reearch wih Applicaion Jul., 218, Vol. 38, No. 4, pp. 418 426 DOI:1.377/j.in:295-2651.218.4.9 Hp://jmre.lu.eu.cn Fracional Brownian Brige Meaure an Their Inegraion by Par Formula

More information

Let. x y. denote a bivariate time series with zero mean.

Let. x y. denote a bivariate time series with zero mean. Linear Filer Le x y : T denoe a bivariae ime erie wih zero mean. Suppoe ha he ime erie {y : T} i conruced a follow: y a x The ime erie {y : T} i aid o be conruced from {x : T} by mean of a Linear Filer.

More information

2. VECTORS. R Vectors are denoted by bold-face characters such as R, V, etc. The magnitude of a vector, such as R, is denoted as R, R, V

2. VECTORS. R Vectors are denoted by bold-face characters such as R, V, etc. The magnitude of a vector, such as R, is denoted as R, R, V ME 352 VETS 2. VETS Vecor algebra form he mahemaical foundaion for kinemaic and dnamic. Geomer of moion i a he hear of boh he kinemaic and dnamic of mechanical em. Vecor anali i he imehonored ool for decribing

More information

On the Benney Lin and Kawahara Equations

On the Benney Lin and Kawahara Equations JOURNAL OF MATEMATICAL ANALYSIS AND APPLICATIONS 11, 13115 1997 ARTICLE NO AY975438 On he BenneyLin and Kawahara Equaion A Biagioni* Deparmen of Mahemaic, UNICAMP, 1381-97, Campina, Brazil and F Linare

More information

Use of variance estimation in the multi-armed bandit problem

Use of variance estimation in the multi-armed bandit problem Ue of variance eimaion in he muli-armed bi problem Jean Yve Audiber CERTIS - Ecole de Pon 19, rue Alfred Nobel - Cié Decare 77455 Marne-la-Vallée - France audiber@cerienpcfr Rémi Muno INRIA Fuur, Grappa

More information

A proof of Ito's formula using a di Title formula. Author(s) Fujita, Takahiko; Kawanishi, Yasuhi. Studia scientiarum mathematicarum H Citation

A proof of Ito's formula using a di Title formula. Author(s) Fujita, Takahiko; Kawanishi, Yasuhi. Studia scientiarum mathematicarum H Citation A proof of Io's formula using a di Tile formula Auhor(s) Fujia, Takahiko; Kawanishi, Yasuhi Sudia scieniarum mahemaicarum H Ciaion 15-134 Issue 8-3 Dae Type Journal Aricle Tex Version auhor URL hp://hdl.handle.ne/186/15878

More information

Network Flows: Introduction & Maximum Flow

Network Flows: Introduction & Maximum Flow CSC 373 - lgorihm Deign, nalyi, and Complexiy Summer 2016 Lalla Mouaadid Nework Flow: Inroducion & Maximum Flow We now urn our aenion o anoher powerful algorihmic echnique: Local Search. In a local earch

More information

BSDE's, Clark-Ocone formula, and Feynman-Kac formula for Lévy processes Nualart, D.; Schoutens, W.

BSDE's, Clark-Ocone formula, and Feynman-Kac formula for Lévy processes Nualart, D.; Schoutens, W. BSD', Clark-Ocone formula, and Feynman-Kac formula for Lévy procee Nualar, D.; Schouen, W. Publihed: 1/1/ Documen Verion Publiher PDF, alo known a Verion of ecord (include final page, iue and volume number)

More information

To become more mathematically correct, Circuit equations are Algebraic Differential equations. from KVL, KCL from the constitutive relationship

To become more mathematically correct, Circuit equations are Algebraic Differential equations. from KVL, KCL from the constitutive relationship Laplace Tranform (Lin & DeCarlo: Ch 3) ENSC30 Elecric Circui II The Laplace ranform i an inegral ranformaion. I ranform: f ( ) F( ) ime variable complex variable From Euler > Lagrange > Laplace. Hence,

More information

Measure-valued Diffusions and Stochastic Equations with Poisson Process 1

Measure-valued Diffusions and Stochastic Equations with Poisson Process 1 Publihed in: Oaka Journal of Mahemaic 41 (24), 3: 727 744 Meaure-valued Diffuion and Sochaic quaion wih Poion Proce 1 Zongfei FU and Zenghu LI 2 Running head: Meaure-valued Diffuion and Sochaic quaion

More information

Physics 240: Worksheet 16 Name

Physics 240: Worksheet 16 Name Phyic 4: Workhee 16 Nae Non-unifor circular oion Each of hee proble involve non-unifor circular oion wih a conan α. (1) Obain each of he equaion of oion for non-unifor circular oion under a conan acceleraion,

More information

CHAPTER 7: SECOND-ORDER CIRCUITS

CHAPTER 7: SECOND-ORDER CIRCUITS EEE5: CI RCUI T THEORY CHAPTER 7: SECOND-ORDER CIRCUITS 7. Inroducion Thi chaper conider circui wih wo orage elemen. Known a econd-order circui becaue heir repone are decribed by differenial equaion ha

More information

REPRESENTATION THEOREMS FOR BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS. BY JIN MA 1 AND JIANFENG ZHANG Purdue University and University of Minnesota

REPRESENTATION THEOREMS FOR BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS. BY JIN MA 1 AND JIANFENG ZHANG Purdue University and University of Minnesota The Annal of Applied Probabiliy 22, Vol. 12, No. 4, 139 1418 REPRESENTATION THEOREMS FOR BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS BY JIN MA 1 AND JIANFENG ZHANG Purdue Univeriy and Univeriy of Minneoa

More information

Stochastic Optimal Control in Infinite Dimensions: Dynamic Programming and HJB Equations

Stochastic Optimal Control in Infinite Dimensions: Dynamic Programming and HJB Equations Sochaic Opimal Conrol in Infinie Dimenion: Dynamic Programming and HJB Equaion G. Fabbri 1 F. Gozzi 2 and A. Świe ch3 wih Chaper 6 by M. Fuhrman 4 and G. Teiore 5 1 Aix-Mareille Univeriy (Aix-Mareille

More information

CHAPTER 7. Definition and Properties. of Laplace Transforms

CHAPTER 7. Definition and Properties. of Laplace Transforms SERIES OF CLSS NOTES FOR 5-6 TO INTRODUCE LINER ND NONLINER PROBLEMS TO ENGINEERS, SCIENTISTS, ND PPLIED MTHEMTICINS DE CLSS NOTES COLLECTION OF HNDOUTS ON SCLR LINER ORDINRY DIFFERENTIL EQUTIONS (ODE")

More information

Chapter 7: Inverse-Response Systems

Chapter 7: Inverse-Response Systems Chaper 7: Invere-Repone Syem Normal Syem Invere-Repone Syem Baic Sar ou in he wrong direcion End up in he original eady-ae gain value Two or more yem wih differen magniude and cale in parallel Main yem

More information

Moral Hazard in Dynamic Risk Management

Moral Hazard in Dynamic Risk Management Moral Hazard in Dynamic Rik Managemen Jakša Cvianić, Dylan Poamaï and Nizar ouzi arxiv:146.5852v2 [q-fin.pm] 16 Mar 215 March 17, 215 Abrac. We conider a conracing problem in which a principal hire an

More information

Backward Stochastic Partial Differential Equations with Jumps and Application to Optimal Control of Random Jump Fields

Backward Stochastic Partial Differential Equations with Jumps and Application to Optimal Control of Random Jump Fields Backward Sochaic Parial Differenial Equaion wih Jump and Applicaion o Opimal Conrol of Random Jump Field Bern Økendal 1,2, Frank Proke 1, Tuheng Zhang 1,3 June 7, 25 Abrac We prove an exience and uniquene

More information

The Asymptotic Behavior of Nonoscillatory Solutions of Some Nonlinear Dynamic Equations on Time Scales

The Asymptotic Behavior of Nonoscillatory Solutions of Some Nonlinear Dynamic Equations on Time Scales Advances in Dynamical Sysems and Applicaions. ISSN 0973-5321 Volume 1 Number 1 (2006, pp. 103 112 c Research India Publicaions hp://www.ripublicaion.com/adsa.hm The Asympoic Behavior of Nonoscillaory Soluions

More information

Approximate Controllability of Fractional Stochastic Perturbed Control Systems Driven by Mixed Fractional Brownian Motion

Approximate Controllability of Fractional Stochastic Perturbed Control Systems Driven by Mixed Fractional Brownian Motion American Journal of Applied Mahemaic and Saiic, 15, Vol. 3, o. 4, 168-176 Available online a hp://pub.ciepub.com/ajam/3/4/7 Science and Educaion Publihing DOI:1.1691/ajam-3-4-7 Approximae Conrollabiliy

More information

Backward stochastic dynamics on a filtered probability space

Backward stochastic dynamics on a filtered probability space Backward sochasic dynamics on a filered probabiliy space Gechun Liang Oxford-Man Insiue, Universiy of Oxford based on join work wih Terry Lyons and Zhongmin Qian Page 1 of 15 gliang@oxford-man.ox.ac.uk

More information

Backward doubly stochastic di erential equations with quadratic growth and applications to quasilinear SPDEs

Backward doubly stochastic di erential equations with quadratic growth and applications to quasilinear SPDEs Backward doubly sochasic di erenial equaions wih quadraic growh and applicaions o quasilinear SPDEs Badreddine MANSOURI (wih K. Bahlali & B. Mezerdi) Universiy of Biskra Algeria La Londe 14 sepember 2007

More information

CSC 364S Notes University of Toronto, Spring, The networks we will consider are directed graphs, where each edge has associated with it

CSC 364S Notes University of Toronto, Spring, The networks we will consider are directed graphs, where each edge has associated with it CSC 36S Noe Univeriy of Torono, Spring, 2003 Flow Algorihm The nework we will conider are direced graph, where each edge ha aociaed wih i a nonnegaive capaciy. The inuiion i ha if edge (u; v) ha capaciy

More information

Additional Methods for Solving DSGE Models

Additional Methods for Solving DSGE Models Addiional Mehod for Solving DSGE Model Karel Meren, Cornell Univeriy Reference King, R. G., Ploer, C. I. & Rebelo, S. T. (1988), Producion, growh and buine cycle: I. he baic neoclaical model, Journal of

More information

Approximation for Option Prices under Uncertain Volatility

Approximation for Option Prices under Uncertain Volatility SIAM J. FINANCIAL MATH. Vol. 5, pp. 360 383 c 014 Sociey for Indurial and Applied Mahemaic Approximaion for Opion Price under Uncerain Volailiy Jean-Pierre Fouque and Bin Ren Abrac. In hi paper, we udy

More information

U T,0. t = X t t T X T. (1)

U T,0. t = X t t T X T. (1) Gauian bridge Dario Gabarra 1, ommi Soinen 2, and Eko Valkeila 3 1 Deparmen of Mahemaic and Saiic, PO Box 68, 14 Univeriy of Helinki,Finland dariogabarra@rnihelinkifi 2 Deparmen of Mahemaic and Saiic,

More information

SINGULAR FORWARD BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS AND EMISSIONS DERIVATIVES

SINGULAR FORWARD BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS AND EMISSIONS DERIVATIVES he Annal of Applied Probabiliy 213, Vol. 23, No. 3, 186 1128 DOI: 1.1214/12-AAP865 Iniue of Mahemaical Saiic, 213 SINGULAR FORWARD BACKWARD SOCHASIC DIFFERENIAL EQUAIONS AND EMISSIONS DERIVAIVES BY RENÉ

More information

LECTURE 1: GENERALIZED RAY KNIGHT THEOREM FOR FINITE MARKOV CHAINS

LECTURE 1: GENERALIZED RAY KNIGHT THEOREM FOR FINITE MARKOV CHAINS LECTURE : GENERALIZED RAY KNIGHT THEOREM FOR FINITE MARKOV CHAINS We will work wih a coninuous ime reversible Markov chain X on a finie conneced sae space, wih generaor Lf(x = y q x,yf(y. (Recall ha q

More information

Quadratic Backward Stochastic Volterra Integral Equations

Quadratic Backward Stochastic Volterra Integral Equations Quadraic Backward Sochaic Volerra Inegral Equaion Hanxiao Wang Jingrui Sun Jiongmin Yong Ocober 25, 218 arxiv:181.1149v1 mah.pr 24 Oc 218 Abrac. Thi paper i concerned wih backward ochaic Volerra inegral

More information

Spectral gaps for the O-U/Stochastic heat processes on path space over a Riemannian manifold with boundary

Spectral gaps for the O-U/Stochastic heat processes on path space over a Riemannian manifold with boundary arxiv:181.186v1 [mah.pr] 5 Dec 18 ec1 Specral gap for he O-U/Sochaic hea procee on pah pace over a Riemannian manifold wih boundary Bo Wu School of Mahemaical Science, Fudan Univeriy, Shanghai 433, China

More information

REFLECTED SOLUTIONS OF BACKWARD SDE S, AND RELATED OBSTACLE PROBLEMS FOR PDE S

REFLECTED SOLUTIONS OF BACKWARD SDE S, AND RELATED OBSTACLE PROBLEMS FOR PDE S The Annals of Probabiliy 1997, Vol. 25, No. 2, 72 737 REFLECTED SOLUTIONS OF BACKWARD SDE S, AND RELATED OBSTACLE PROBLEMS FOR PDE S By N. El Karoui, C. Kapoudjian, E. Pardoux, S. Peng and M. C. Quenez

More information

On a Fractional Stochastic Landau-Ginzburg Equation

On a Fractional Stochastic Landau-Ginzburg Equation Applied Mahemaical Sciences, Vol. 4, 1, no. 7, 317-35 On a Fracional Sochasic Landau-Ginzburg Equaion Nguyen Tien Dung Deparmen of Mahemaics, FPT Universiy 15B Pham Hung Sree, Hanoi, Vienam dungn@fp.edu.vn

More information

arxiv:math/ v1 [math.pr] 6 Feb 2007

arxiv:math/ v1 [math.pr] 6 Feb 2007 Sochaic Differenial Game and Vicoiy Soluion of Hamilon-Jacobi-Bellman-Iaac Equaion arxiv:mah/0702131v1 [mah.pr] 6 Feb 2007 Rainer Buckdahn Déparemen de Mahémaique, Univerié de Breagne Occidenale, 6, avenue

More information

Main Reference: Sections in CLRS.

Main Reference: Sections in CLRS. Maximum Flow Reied 09/09/200 Main Reference: Secion 26.-26. in CLRS. Inroducion Definiion Muli-Source Muli-Sink The Ford-Fulkeron Mehod Reidual Nework Augmening Pah The Max-Flow Min-Cu Theorem The Edmond-Karp

More information

Network Flow. Data Structures and Algorithms Andrei Bulatov

Network Flow. Data Structures and Algorithms Andrei Bulatov Nework Flow Daa Srucure and Algorihm Andrei Bulao Algorihm Nework Flow 24-2 Flow Nework Think of a graph a yem of pipe We ue hi yem o pump waer from he ource o ink Eery pipe/edge ha limied capaciy Flow

More information

Problem Set If all directed edges in a network have distinct capacities, then there is a unique maximum flow.

Problem Set If all directed edges in a network have distinct capacities, then there is a unique maximum flow. CSE 202: Deign and Analyi of Algorihm Winer 2013 Problem Se 3 Inrucor: Kamalika Chaudhuri Due on: Tue. Feb 26, 2013 Inrucion For your proof, you may ue any lower bound, algorihm or daa rucure from he ex

More information

Hamilton- J acobi Equation: Explicit Formulas In this lecture we try to apply the method of characteristics to the Hamilton-Jacobi equation: u t

Hamilton- J acobi Equation: Explicit Formulas In this lecture we try to apply the method of characteristics to the Hamilton-Jacobi equation: u t M ah 5 2 7 Fall 2 0 0 9 L ecure 1 0 O c. 7, 2 0 0 9 Hamilon- J acobi Equaion: Explici Formulas In his lecure we ry o apply he mehod of characerisics o he Hamilon-Jacobi equaion: u + H D u, x = 0 in R n

More information

Simulation of BSDEs and. Wiener Chaos Expansions

Simulation of BSDEs and. Wiener Chaos Expansions Simulaion of BSDEs and Wiener Chaos Expansions Philippe Briand Céline Labar LAMA UMR 5127, Universié de Savoie, France hp://www.lama.univ-savoie.fr/ Workshop on BSDEs Rennes, May 22-24, 213 Inroducion

More information

Physics 235 Chapter 2. Chapter 2 Newtonian Mechanics Single Particle

Physics 235 Chapter 2. Chapter 2 Newtonian Mechanics Single Particle Chaper 2 Newonian Mechanics Single Paricle In his Chaper we will review wha Newon s laws of mechanics ell us abou he moion of a single paricle. Newon s laws are only valid in suiable reference frames,

More information

arxiv: v1 [math.pr] 23 Apr 2018

arxiv: v1 [math.pr] 23 Apr 2018 arxiv:184.8469v1 [mah.pr] 23 Apr 218 The Neumann Boundary Problem for Ellipic Parial ifferenial Equaion wih Nonlinear ivergence Term Xue YANG Tianjin Univeriy e-mail: xyang213@ju.edu.cn Jing ZHANG Fudan

More information

arxiv: v1 [math.pr] 19 Feb 2011

arxiv: v1 [math.pr] 19 Feb 2011 A NOTE ON FELLER SEMIGROUPS AND RESOLVENTS VADIM KOSTRYKIN, JÜRGEN POTTHOFF, AND ROBERT SCHRADER ABSTRACT. Various equivalen condiions for a semigroup or a resolven generaed by a Markov process o be of

More information

Dynamic Systems and Applications 16 (2007) BACKWARD STOCHASTIC DIFFERENTIAL INCLUSIONS 1. INTRODUCTION. x t + t

Dynamic Systems and Applications 16 (2007) BACKWARD STOCHASTIC DIFFERENTIAL INCLUSIONS 1. INTRODUCTION. x t + t Dynamic Syem and Applicaion 16 (7) 11-14 BACKWARD STOCHASTIC DIFFERENTIAL INCLUSIONS MICHA L KISIELEWICZ Faculy of Mahemaic, Compuer Science and Economeric, Univeriy of Zielona Góra, Podgórna 5, 65 46

More information

Existence of Stepanov-like Square-mean Pseudo Almost Automorphic Solutions to Nonautonomous Stochastic Functional Evolution Equations

Existence of Stepanov-like Square-mean Pseudo Almost Automorphic Solutions to Nonautonomous Stochastic Functional Evolution Equations Exience of Sepanov-like Square-mean Peudo Almo Auomorphic Soluion o Nonauonomou Sochaic Funcional Evoluion Equaion Zuomao Yan Abrac We inroduce he concep of bi-quare-mean almo auomorphic funcion and Sepanov-like

More information

L p -Solutions for Reflected Backward Stochastic Differential Equations

L p -Solutions for Reflected Backward Stochastic Differential Equations L -Soluion for Refleced Backward Sochaic Differenial Equaion Saïd Hamadène and Alexandre Poier Laboraoire de aiique e roceu Univerié du Maine Avenue Olivier Meiaen F-7285 Le Man Cedex 9 France March 27,

More information

and Applications Alexander Steinicke University of Graz Vienna Seminar in Mathematical Finance and Probability,

and Applications Alexander Steinicke University of Graz Vienna Seminar in Mathematical Finance and Probability, Backward Sochasic Differenial Equaions and Applicaions Alexander Seinicke Universiy of Graz Vienna Seminar in Mahemaical Finance and Probabiliy, 6-20-2017 1 / 31 1 Wha is a BSDE? SDEs - he differenial

More information

Time Varying Multiserver Queues. W. A. Massey. Murray Hill, NJ Abstract

Time Varying Multiserver Queues. W. A. Massey. Murray Hill, NJ Abstract Waiing Time Aympoic for Time Varying Mulierver ueue wih Abonmen Rerial A. Melbaum Technion Iniue Haifa, 3 ISRAEL avim@x.echnion.ac.il M. I. Reiman Bell Lab, Lucen Technologie Murray Hill, NJ 7974 U.S.A.

More information

arxiv: v2 [math.oc] 15 Feb 2019

arxiv: v2 [math.oc] 15 Feb 2019 OPTIMAL CONTROL OF DEBT-TO-GDP RATIO IN AN N-STATE REGIME SWITCHING ECONOMY GIORGIO FERRARI, NEOFYTOS RODOSTHENOUS arxiv:188.1499v2 mah.oc 15 Feb 219 Abrac. We olve an infinie ime-horizon bounded-variaion

More information

An ergodic BSDE approach to forward entropic risk measures: representation and large-maturity behavior

An ergodic BSDE approach to forward entropic risk measures: representation and large-maturity behavior An ergodic BSDE approach o forward enropic rik meaure: repreenaion and large-mauriy behavior W. F. Chong Y. Hu G. Liang. Zariphopoulou Fir verion: January 2016; hi verion: November 30, 2016 Abrac Uing

More information

Energy Equality and Uniqueness of Weak Solutions to MHD Equations in L (0,T; L n (Ω))

Energy Equality and Uniqueness of Weak Solutions to MHD Equations in L (0,T; L n (Ω)) Aca Mahemaica Sinica, Englih Serie May, 29, Vol. 25, No. 5, pp. 83 814 Publihed online: April 25, 29 DOI: 1.17/1114-9-7214-8 Hp://www.AcaMah.com Aca Mahemaica Sinica, Englih Serie The Ediorial Office of

More information

Cash Flow Valuation Mode Lin Discrete Time

Cash Flow Valuation Mode Lin Discrete Time IOSR Journal of Mahemaics (IOSR-JM) e-issn: 2278-5728,p-ISSN: 2319-765X, 6, Issue 6 (May. - Jun. 2013), PP 35-41 Cash Flow Valuaion Mode Lin Discree Time Olayiwola. M. A. and Oni, N. O. Deparmen of Mahemaics

More information