首页 > 范文大全 > 正文

基于指数障碍期权的抛物方程的存在性和唯一性

开篇:润墨网以专业的文秘视角,为您筛选了一篇基于指数障碍期权的抛物方程的存在性和唯一性范文,如需获取更多写作素材,在线客服老师一对一协助。欢迎您的阅读与分享!

摘要考虑了一类基于指数障碍期权的拟线性抛物方程. 首先在b(t,x)=c(t,x)=0情形下运用标准的Schauder理论证明了该抛物型方程问题存在一个属于Cα,1+α/2的唯一解. 其次, 运用变换的方法将该结论推广到了一般方程.

关键词修正的BlackScholes 方程; 指数障碍期权; Schauder估计; 存在性; 唯一性

中图分类号O175.2, F830.9 文献标识码A

Appendix

(A) f0(t)satisfies the following properties:

f0(t)≤1λfL

SymboleB@ (V),

f′0(t)≤Ct-2+αfCα,α/2(),

f0(t)-f0(s)≤CfCα,α/2()t-sα/2,

f(x,t)-a(x,t)f0(t)≤CfCα,α/2().

Where C depends on λ,μ,TandaCα,α/2().

Proof Note that f0(t)=1t2∫tt-t2f(0,τ)a(0,τ)dτ, according to (19), one gets

f0(t)≤1λt2∫tt-t2f(0,τ)dτ≤1λt2fL

SymboleB@ (V)·t2=1λfL

SymboleB@ (V).

Hence the first inequality is obtained. Defining

h(τ)=f(0,τ)a(0,τ).

One gets

f0(t)=1t2∫tt-t2h(τ)dτ,

Such that

f′0(t)=I1+I2+I3, (64)

where

I1=2t3∫tt-t2h(τ)dτ≤2t3hL

SymboleB@ (V)t2≤hL

SymboleB@ (V)t-2+α, (65)

I2=1t2h(t)-h(t-t2)≤CfCα,α/2()·t-2+α, (66)

and

I3=1t2h(t-t2)2t

SymbolcB@ CfL

SymboleB@ (V)t-2+α, (67)

Substituting(65)-(67)into(64), hence the second inequality is hold. Next we prove the third inequality. Since

f0(t)=1t2∫tt-t2h(τ)dτ=1t2∫10h(t-λt2)dλ, (68)

By (68), it follows that

f0(t)-f0(s)≤CfCα,α/2()t-sα2.

Finally, we prove the last inequality. We have

f(x,t)-a(x,t)f0(t)=J1+J2+J3,

where

J1=f(x,t)-f(0,t)≤fCα,α/2()xα≤fCα,α/2()tα,(69)

J2≤μ1t2∫tt-t2hCα/20,T(t-τ)α/2dτ≤CfCα,α/2()tα, (70)

J3=a(0,t)-a(x,t)·f0(t)≤CfL

SymboleB@ (V)tα. (71)

Taking into account (69)-(71) and the triangle inequality, the conclusion is obtained.

(B) v(x,t)∈C2+α,1+α/2(); moreover

v(x,t)C2+α,1+α/2()≤CfCα,1+α/2().

Proof According to the definition ofv(x,t), one gets

xv(x,t)=-xf0(t), 2x2v(x,t)=-f0(t),

tv(x,t)=12f′0(t)(t2-x2)+tf0(t).

Now we provexv∈C1+α,(1+α)/2(). In fact for fixed x≠0,t1>t2≥0, ift1-t2≤x, then using the second inequality of (A), it follows that

xf0(t1)-xf0(t2)=xf′0(ξ)·t1-t2≤CfCα,α/2()t1-t21+α2.

Ift1-t2>x, we also can find that

xf0(t1)-xf0(t2)

≤CfCα,α/2()xt1-t2α2

≤CfCα,α/2()t1-t21+α2,

This means thatxv∈C1+α,(1+α)/2(). By using similar way more or less, we can prove tv∈Cα,α/2(). It is clear that2x2v∈Cα,α/2().We have completed the proof1Introduction

Nowadays there has been an increasing interest in problems arising in Financial Mathematics and in particular on option pricing. The standard approach to this problem leads to the study of equations of a parabolic type.

Now, we focus on the exponential double barrier options. A double barrier options has two barriers, upstream barrier at U(t)and downstream barrier L(t)which dependent on t. In its simplest form, when the asset price reaches one of the two barriers either nullification (knockout) or activation (knockin) of the option contract is triggered. A more complicate payoff structure may include a rebate payment upon hitting either one of the knockout barriers as compensation to the option holder.

At the beginning of our main aim, let us briefly describe the model as follows. Let t be the time and St be the price of stock. Consider a derivative security whose price say V depend on St and t. Assume that there is a riskfree bond which earns a riskfree rater, satisfying the relation that dBt=rBtdt. In addition, we assume that the stock pricing St follows the geometric Brownian motion.

dSt=μStdt+σ(St)dWHt,

where WHt is the fractional Brownian motion with Hurst parameter H. ∫t0σ(St)dWHt is WickItSkorohod integral not RiemanStieltjes product, for detail see[1-3]. σ(St)is a volatility function that may depend on the level of the asset price(stock price). By the selffinancing strategy[3] and It formula[4] for fractional Brownian motion, the exponential double barrier call option premiumV=V(S,t,K,T) satisfies the following generalized BlackScholes model

Vt+Hσ2t2H-1(S)SVSS+rSVS-rV=f(S,t), (1)

V(T,S)=max{S-K,0} (terminal payoff),(2)

V(L(t),t)=g1(t)V(U(t),t)=g2(t)(knockout condition) (3)

Where T is the maturity, K is the strike. f(S,t)is the cash flow. g1(t)(g2(t)) denote the timedependent rebate paid to the holder when the lower barrier ( upper barrier) is hit .

When σ2(S)is a constant,f(S,t)=0, the partial differential equation (1) becomes the well known BS equation[5]. In this derivation, we replicated the derivative with a stock and a bond. However, it is also possible to derive the BS equation by replicating the bond with a stock and a derivative, or by replicating the stock with a bond and a derivative. Moreover, as an interesting consequence of noarbitrage arguments (see[6]), the drift rate does not enter in (1). In general for European options, the BS equation results in a boundary value problem of a diffusion equation, whereas for American options the BS equation results in a free boundary value problem (see[7] and[8]). There are usually two ways to solve these kinds of option pricing problem: the analytic and numerical approaches (see \[9\], \[10\], \[11\] and \[12\]).

We rewrite the L(t),U(t),

L(t)=L0exp{s1(t)}, U(t)=U0exp{s2(t)},

s1(t)=s2(t)=0, s′1(t)

If we assume the special caseL(T)=U(T), according to[5], the option can not be execution, because the stock price will hit the barrier soon or later. Hence, They can not gain any profit at time T, but they can still gain the rebate paid when the barrier is hit. By introducing the change of variables x=lnS,τ=T-t, then (1)-(3) becomes

Vτ-Hσ2(ex)t2H-1Vxx+Hσ2(ex)t2H-1-rVx+rV=f(ex,t),(4)

V(s1(t),t)=g1(t),(5)

V(s2(t),t)=g2(t) (s1(0)=s2(0)) . (6)

The rest of the paper is organized as follows. In Section 2, we introduce the hypotheses and our main result. In section 3, we prove our main result under the conditions s1(t)=t,s2(t)=-t.Section 4 contain the proof of theorem 1.

2Hypotheses and main result

The (4)~(6) motivate us to consider the following parabolic model for the exponential double barrier options

tu-a(t,x)2x2u+b(t,x)xu+c(t,x)u

=f(t,x),(7)

u(t,s1(t))=g1(t), (8)

u(t,s2(t))=g2(t). (9)

Before starting, let us assume that

a,b,c,f∈Cα,α/2(Ω),0

SymboleB@ ,(11)

s1(t),s2(t)∈C1+α/2(0,T),

s1(0)=s2(0)=h,s′1(t)

g(x,t)∈C2+α,1+α/2(Ω) satisfying

g(si(t),t)=si(t),i=1,2. (13)

The main result is described as follows.

Theorem 1Under the assumptions (10)~(13), the problem (7)~(9) has a unique solution u∈C2+α,1+α/2(Ω) which satisfies

uC2+α,1+α/2(Ω)≤CfCα,1+α/2(Ω)+gC2+α,1+α/2(Ω),(14)

where C depends on Cα,α/2(Ω) norms of a,b,c, C1+α/2(0,T)norms of s1(t),s2(t), and λ,μ.

Remark 1In view of condition (12) we only need to consider s1(0)=s2(0)=0, otherwise we assume that x′=x-h, (8) and (9) can be changed as follow

u(t,s1(t)-h)=g1(t)u(t,s2(t)-h)=g2(t).

Remark 2In view of condition (13) we only need to consider homogeneous boundary conditions, otherwise we assume that

u1=u-x-s1(t)g2(t)-x-s2(t)g1(t)s2(t)-s1(t). (15)

It follows that

tu-a(t,x)2x2u+b(t,x)xu+c(t,x)u=f(t,x)+g2(t)-g1(t)s2(t)-s1(t)

+x-s1(t)g′2(t)-x-s2(t)g′1(t)+s′2(t)g1(t)-s′1(t)g2(t)s2(t)-s1(t)

-x-s1(t)g2(t)-x-s2(t)g1(t)s′2(t)-s′1(t)s2(t)-s1(t),

u(t,s1(t))=0, u(t,s2(t))=0.

So from now on, we assumeg0,h=0.

3A priori estimate of C2+α,1+α/2type

Firstly, Assuming that b(x,t)=c(x,t)=0, s1(t)=t,s2(t)=-t in (7), (8) and (9), one gets

tu-a(t,x)2x2u=f(t,x),x

u=0, on x=±t. (17)

Let us defining

V={(x,t)x

a(x,t)∈Cα,α/2(), (18)

SymboleB@ ,(19)

where λ and μ are constants. Hence the following theorem can be hold.

Theorem 2Under the assumptions (18) and (19), assumeu∈C()∩C2+α,1+α/2(V)satisfying (16) and (17); then u∈C2+α,1+α/2()and

uC2+α,1+α/2()≤CfCα,α/2(),(20)

where C depends on λ,μ,T andaCα,α/2().

Assume that

f0(t)=1t2∫tt-t2f(0,τ)a(0,τ)dτ,

v(x,t)=12f0(t)(t2-x2),

w(x,t)=u(x,t)-v(x,t). (21)

Hence, one gets

tw-a(x,t)2x2w=g(x,t)x

w=0 on x=±t,(23)

where

g(x,t)=f(x,t)-12f′0(t)(t2-x2)-tf0(t)-a(x,t)f0(t).

Taking into account appendix (A), one gets

g(x,t)≤CfCα,α/2()tα, (24)

g(x,t)Cα,α/2()≤CfCα,α/2().(25)

Before we proof Theorem 2, we need the following lemma.

Lemma 1

w(x,t)

SymbolcB@ CfCα,α/2()tα(t2-x2),

x

Proof Note thatx

SymboleB@ , one gets

αtα-1(t2-x2)+2t1+α+2a(x,t)tα

≥2t1+α+2a(x,t)tα≥2t1+α+2λtα

=2λtα(1+t)≥2λtα. (27)

Using (19), one gets

tCfCα,α/2()tα(t2-x2)+w

-a(x,t)2t2CfCα,α/2()tα(t2-x2)+w

=CfCα,α/2()αtα-1(t2-x2)+2t1+α+2a(x,t)tα

+g(x,t)≥CfCα,α/2()tα+g≥0,and

tCfCα,α/2()tα(t2-x2)-w

-a(x,t)2t2CfCα,α/2()tα(t2-x2)-w

=CfCα,α/2()αtα-1(t2-x2)+2t1+α+2a(x,t)tα

-g(x,t)≥CfCα,α/2()tα-g≥0,

and so (26) is true by the maximum principle.

Proof of theorem 2

For any0

Qr={(x,t)x

Q+r={(x,t)x

Q-r={(x,t)x

Using a scaling technique, let

x=ry,t=r+2r2(τ-1), (28)

and then QrQr+Qr-are transformed to Q,Q+,Q-, where

Q={(y,τ)y

Q-={(y,τ)y

Defining the transformation

W(y,τ)=w(x,t),A(y,τ)=a(x,t),

G(y,τ)=g(x,t),

the (22)-(23) can be written with

τW(y,τ)-2A(y,τ)2y2W(y,τ)

=2r2G(y,τ),

W=0 on y=1+2r(τ-1).

Using the standard Schauder theory of parabolic equations (see \[13-15\]), and taking the inverse transformation of (28) we have

r2+α2x2wCα,α/2(r+)+r2+ατwCα,α/2(Q-r+)

+r2+αxwCt(α+1)/2(r+)+r22x2wL

SymboleB@ (r+)

+r2twL

SymboleB@ (r+)+rxwL

SymboleB@ (Q-r+)

≤Cr2+αg(x,t)Cα,α/2(r)+r2g(x,t)L

SymboleB@ (Qr)

+w(x,t)L

SymboleB@ (Qr). (29)

Substituting (24)-(26) into (29) , it follows that

2x2wCα,α/2(r+)+twCα,α/2(r+)

+xwCt(α+1)/2(r+)+r-α2x2wL

SymboleB@ (Qr+)

+r-αtwL

SymboleB@ (Qr+)+r-1-αxwL

SymboleB@ (Qr+)

≤CfCα,α/2() (30)

For any (x1,t1),(x2,t2)∈,t1≥t2, let r=t1 in (30). If(x1-x2)2+t1-t2≤r, then using the second inequality of appendix (A), one gets

2x2w(x1,t1)-2x2w(x2,t2)

+tw(x1,t1)-tw(x2,t2)

≤CfCα,α/2()(x1-x2)2+t1-t2α, (31)

xw(x,t1)-xw(x,t2)

≤CfCα,α2()t1-t2(1+α)/2. (32)

If (x1-x2)2+t1-t2>r, then also using the second inequality of appendix (A) , it follows that

2x2w(x1,t1)-2x2w(x2,t2)

+tw(x1,t1)-tw(x2,t2)

≤CfCα,α/2()(x1-x2)2+t1-t2α, (33)

xw(x,t1)-xw(x,t2)≤xw(x,t1)

+xw(x,t2)≤CfCα,α/2()r(1+α)/2

≤CfCα,α/2()t1-t1(1+α)/2. (34)

Taking into account (30)-(34), it follows that

2x2wCα,α/2()+twCα,α/2()+xwCt(α+1)/2()

+2x2wL

SymboleB@ (V)+twL

SymboleB@ (V)+xwL

SymboleB@ (V)

≤CfCα,α/2()(35)

Combining appendix (B) and (34) we derive the estimates (20). This completes the proof of Theorem2.

Theorem 3The problem (16) and (17) has a unique solutionu∈C2+α,1+α/2()satisfying the estimates (20).

Proof First we consider an approximation problem for problem (16) and (17):

tuε=a(t,x)2x2uε=f(t,x),

x

uε=0 , on x=±t, ε

uε(x,ε)=0,-ε

This problem has a unique solution uε∈C2+α,1+α/2(Vε)∩C(ε), where Vε=V∩{t>ε}; moreover

uεC2+α,1+α/2(2ε)

SymbolcB@ CεfCα,α/2(), (39)

where Cε depends onε.From (39) we know that there is a subsequence of uε(still denoted by uε) and a function u(x,t)∈C2+α,1+α/2(V)∩C(\(0,0)), such that for any τ>0, uεu in C2+α,1+α/2(Vτ), and moreover

tu-a(t,x)2x2u=f(t,x),

x

u=0 , on x=±t, 0

It is clear thatu

SymbolcB@ tfL

SymboleB@ ; it follows thatlim(x,t)(0,0)u(x,t)=0, which means that u(x,t)∈C2+α,1+α/2(V)∩C(). Using Theorem 2 we complete the proof of Theorem 3.2.

Now we consider the problem

tu-a(t,x)2x2u+b(x,t)xu

+c(x,t)u=f(t,x),

x

u=0on x=±t. (41)

On the basis of Theorem 3 and a fixed point technique we can prove the following theorem:

Theorem 4The problem (40) and (42) has a unique solution u∈C2+α,1+α/2() satisfying the estimates(20).

5 The proof of Theorem 1

In this section we consider the problem (7), (8) and (9) (note thatgi(t)0,i=1,2,h=0). Apply a transformation of variables

y=t2x-s1(t)-s2(t)s1(t)-s2(t), t=t.

Defining u(x,t)=v(y,t); it follows that

xu=2ts1(t)-s2(t)yv, (42)

2x2u=2ts1(t)-s2(t)22y2v,(43)

tu=tv+yv·ty(y,t).(44)

The problem (7), (8) and (9) becomes

tv-(y,t)2ts1(t)-s2(t)22y2v

+ty+(y,t)2ts2(t)-s1(t)yv

+(y,t)v=(y,t),(45)

v=0on x=±t , (46)

where

(y,t)=a(x,t), (y,t)=b(x,t),

(y,t)=c(x,t), (y,t)=f(x,t),

x=y2t(s2(t)-s1(t))+12(s2(t)+s1(t)).

Lemma 2

2t/(s1(t)-s2(t))≥C0>0,(47)

2t/(s1(t)-s2(t))∈Cα/2[0,T], (48)

(y,t),(y,t),(y,t),(y,t)∈Cα,α/2(),(49)

ty∈Cα,α/2().(50)

Proof Let us defines(t)=s1(t)-s2(t); then, by (12),

s(0)=0, s′(0)>0, s(t)∈C1+α/2[0,T],

s(t)>0(51)

for all 0

limt0+2t/s(t)=2/s′(0)>0,

there is a C0>0, such that 2t/s(t)≥C0.Next we prove (48). Considerings(t)=t∫10s′(λt)dλ, lettingC1=s′(0)/2>0, such that

s(t)t=∫10s′(λt)dλ≥s′(0)>C1>0, (52)

and therefore

2t1s(t1)-2t2s(t2)=2∫10s′(λt1)dλ-∫10s′(λt2)dλ∫10s′(λt1)dλ∫10s′(λt2)dλ,

and hence

2t1s(t1)-2t2s(t2)≤2C21∫10s′(λt1)-s′(λt2)dλ

≤2C21Ct1-t2α/2.

This completes the proof of (48). Next we prove (49). In fact

(y1,t)-(y2,t)y1-y2α≤aCα,α/2(Ω)s(t)2tα

≤CaCα,α/2(Ω).

On the other hand,

(y1,t)-(y2,t)≤|a(y2t1s(t1)+12(s1(t1)

+s2(t2)),t1)-a(y2t1s(t1)+12(s1(t1)

+s2(t1)),t2)+|a(y2t1s(t1)+12(s1(t1)

+s2(t2)),t2)-a(y2t2s(t2)+12(s1(t2)

+s2(t2)),t2)=defI1+I2

where

I1≤aCα/2([0,T])t1-t2α/2

≤aCα,α/2(Ω)t1-t2α/2.

By using (47) and (52), one gets

y2t1s(t1)-y2t2s(t2)≤Ct1-t2. (53)

It is clear that

12s1(t1)+s2(t1)-12s1(t2)+s2(t2)

By (53)-(54) , it follows that

I2≤CaCα,α/2(Ω)t1-t2α/2,

which means (y,t)∈Cα,α/2(). In the same way we can prove (y,t),(y,t),(y,t)∈Cα,α/2(); in particular,

Cα,α/2()≤Cf(x,t)Cα,α/2(Ω).(55)

Finally we prove (50). Since

ty=yts(t)-ts′(t)s(t)-ts(t)s′1(t)+s′2(t),(56)

it is clear, by (48)and (12), that the second term in the right hand side of (56) is in Cα/2[0,T] and the first term in the right hand side of (56) is (denote it by h(y,t) )

h(y,t)=yt∫10s′(λt)-s′(t)dλ∫10s′(λt)dλ.

Since∫10s′(λt)-s′(t)dλ≤∫10λ-1α/2tα/2dλ=C·tα/2, by using (52), it follows that

h(y1,t)-h(y2,t)≤1C1y1-y2t1-α2

≤1C1y1-y2α2. (57)

On the other hand,

h(y,t1)-h(y,t2)

=yt1-t2t1t2∫10s′(λt1)-s′(t1)dλ∫10s′(λt1)dλ

+|y|t2∫10s′(λt1)-s′(t1)dλ∫10s′(λt1)-s′(t2)dλ∫10s′(λt1)dλ∫10s′(λt2)dλ

+yt2∫10s′(λt1)-s′(t1)-s′(λt2)+s′(t2)dλ∫10s′(λt2)dλ

≤Ct1-t2α/2

and we have completed the proof of Lemma 2.

Lemma 3 ty(y,t)·tv(y,t)∈Cα,α/2() and

ty(y,t)·tv(y,t)Cα,α/2()

≤Cf-(y,t)Cα,α/2(). (58)

Proof According to Theorem 3 and (50), we have that the problem (45) and (46) has a unique solutionv(y,t)∈C2+α/2,1+α/4(); in particular, yv∈C1+α/2,1/2+α/4()and

y(y,t)v(y,t)C1+α/2,1/2+α/4()

≤Cf-(y,t)Cα/2,α/4(). (59)

Using yv(0,0)=0 and (55), one gets

t(y1,t)yv(y1,t)-t(y2,t)yv(y2,t)

≤Cy1-y2t1-α/2y1+Cy1-y2t1-α/2t12+α4

+Cy1-y2

We also can find that

ty(y,t)·tv(y,t)Ctα/2()

≤Cyv(y,t)Ctα/2(),

and this completes the proof of Lemma 3.

Proof of Theorem 1 On the basis of Theorem 4 and (47)-(50), we have that the problem (45) and (46) has a unique solution v(y,t)∈C2+α/2,1+α/4(), and then we rewrite the (45) as

tv-a-(y,t)2ts1(t)-s2(t)22y2v

+(y,t)2ts2(t)-s1(t)yv

+(y,t)v=(y,t)-tyyv,

y

Using (58) and Theorem 4 to the problem (60) and (46), we know that v(y,t)∈C2+α,1+α/2().It is deduced from (43) and (44) and Lemma 2 that

2x2u(x,t),tu(x,t)∈C2+α,1+α/2(Ω).

Now we provexu(x,t)∈Ct1+α2(Ω). By (42),

2t1s(t1)yv(y,t1)-2t2s(t2)yv(y,t2)

≤2t1s(t1)-2t2s(t2)yv(y,t1)-yv(0,t1)

+2t1s(t1)-2t2s(t2)yv(0,t1)-yv(0,0)

+2t2s(t2)yv(y,t1)-yv(y,t2)

=I1+I2+I3,

and

I1≤Cξ1-α/2t1-t2y≤Cξα/2t1-t2

≤Ct1-t2,(61)

where t2>ξ>t1≥y, if t1-t2

I2≤Cξ1-α/2t1-t2t11+α/2

≤Cξ1-α/2t1-t21+α/2t1

≤Ct1-t21+α/2.

Ift1-t2≥t1, applying (48),

I2≤Ct1-t2α/2·t(α+1)/2≤Ct1-t2(α+1)/2,(62)

and, moreover,

I3≤Ct1-t2(α+1)/2. (63)

Combing (61) ,(62)and (63), it follows that

2ts(t)yv(y,t)∈Ct1+α2().

We have completed the proof of Theorem 1.

6 Conclusion

Using the standard Schauder theory, an parabolic problem of determining the market price of exponential double barrier options is discussed in this paper. We prove that the main problem has a unique solution u∈C2+α,1+α/2(Ω)on an irregular Wedge-shaped domain. We also provide a estimate (see [14]).

The parabolic problem discussed in the paper is a very fundamental and important problem in finance, especially in the area of derivative pricing. Based on the rigorous theoretical analysis, the results obtained in the paper are interesting and useful. People who engage in option pricing may apply the method to reconstruct the market price of risk.

References

[1]R J ELLIOTT, J HOEK. A general fractional white noise theory and applications to finance\[J\]. Mathematical Finance, 2003,13 (2) : 301-330.

\[2\]Y HU, B KSENDAL. Fractional white noise calculus and applications to finance, infinite dimensional analysis\[J\]. Quantum Probability and Related Topics, 2003, 6 (1): 1-32 .

\[3\]P GUASONI. No arbitrage under transaction costs, with fractional Brownian motion and beyond\[J\]. Mathematical Finance, 2006, 16(3) : 569-582.

\[4\]F BIAGINI, Y HU, B OKSENDAL,et al. Stochastic calculus for fractional brownian motion and applications\[M\]. New York: Springer, 2008.

\[5\]F BLACK, M SCHOLES. The pricing of options and corporate liabilities\[J\]. Political Economy, 1973, 81(3): 637-654.

\[6\]P WILMOTT, S HOWISON, J DEWYNNE. The mathematics of financial derivatives\[M\]. Cambridge: Cambridge University Press, 1995.

[7]C L GAO, F H YI. A Parabolic variational inequality arising from the valuation of American interest rate options[J]. Applied Mathematics A Journal of Chinese University, 2008, 23(1): 13-22.

[8]F H YI, T YU. Free boundary problem arising from the valuation of the russian option[J]. Acta Mathematicae Applicatae Sinica Chinese Series, 2008, 31(6): 993-1012.

[9]Y D SUN, Y M SHI. European pricing options on modified BlackScholes model [J]. Applied Mathematics A Journal of Chinese University, 2012, 27(1): 23-32.

[10]Y D SUN, Y M SHI. European pricing options on modified BlackScholes model [J]. Journal of Quantitive Economics, 2011, 28(1): 49-51.

\[11\]B DURING, A JUNGEL. Existence and uniqueness of solutions to a quasilinear parabolic equation with quadratic gradients in financial markets\[J\]. Nonlinear Analysis, 2005, 62(3):519–544.

\[12\]R KANGRO, R NICOLAIDES. Far field boundary conditions for Black–Scholes equations\[J\]. SIAM Journal on Numerical Analysis, 2000, 38 (4):1357–1368.

\[13\]G M LIEBERMAN. Second Order Parabolic Differential Equations\[M\]. London:World Scientific Publishing, 1996.

\[14\]O A LADYZENSKAJA, V A SOLONNIKOV. N N Ural’ceva, Linear and Quasilinear Equations of Parabolic Type\[M\]. Chicago: American Mathematical Society, Providence, 1968.

\[15\]L C EVANS. Partial differential equations \[M\]//Graduate Studies in Mathematics, Chicago: American Mathematical Society, 2010.