Abstract
In this paper, we use a generalized Brownian motion process to define an analytic operator-valued Feynman integral. We then establish the existence of the analytic operator-valued generalized Feynman integral. We next investigate a stability theorem for the analytic operator-valued generalized Feynman integral.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Cameron and Storvick [1] introduced an analytic operator-valued function space integral and showed that the integral satisfied an integral equation related to the Schrödinger equation. The existence of this integral was established as an operator from \(L_2(\mathbb R)\) to \(L_2(\mathbb R)\). Since then, Johnson and Lapidus [8] established the existence of the operator-valued function space integral as a bounded linear operator on \(L_2(\mathbb R^n)\) for certain functionals which only define finite Borel measures on the compact interval [0, T] in \(\mathbb R\). These integrals are based on the Wiener integral associated with the Wiener process.
On the other hand, Johnson [7] studied a bounded convergence theorem (stability theorem) for the operator-valued Feynman integral of functionals of the form \(F(x)=\exp \{\int _{0}^{T}\theta (x(s))\hbox {d}s\}\). Chang et al. [2] established a stability theorems for the operator-valued Feynman integral of certain functionals involving some Borel measures on an interval (0, T) as a bounded linear operator from \(L_1(\mathbb R)\) to \(C_0(\mathbb R)\). Moreover, Chang and Lee [5] studied an analytic operator-valued generalized Feynman integral. The integral investigated in [5] is based on the function space integral associated with a generalized Brownian motion process.
The function space \( C_{a,b}[0,T]\) induced by a generalized Brownian motion was introduced by Yeh in [10] and was studied extensively in [3, 4, 6]. In this paper, we define an analytic operator-valued generalized Feynman integral on the function space \(C_{a,b}[0,T]\). We then establish the existence of the analytic operator-valued generalized Feynman integral and investigate a stability theorem for the analytic operator-valued generalized Feynman integral.
2 Definitions and Preliminaries
Let \( D =[0,T]\) and let \((\Omega , {\mathcal {B}},P)\) be a probability measure space. A real-valued stochastic process Y on \((\Omega , {\mathcal {B}}, P)\) and D is called a generalized Brownian motion process if \(Y(0,\omega )=0\) almost everywhere and for \(0=t_0<t_1< \cdots <t_n \le T \), the n-dimensional random vector \((Y(t_1,\omega ), \ldots ,Y(t_n,\omega ))\) is normally distributed with density function
where \(\vec \eta = (\eta _1,\ldots ,\eta _n)\), \(\eta _0=0\), \(\vec t = (t_1,\ldots , t_n)\), a(t) is an absolutely continuous real-valued function on [0, T] with \(a(0)=0\), \(a'(t) \in L^2[0,T]\), and b(t) is a strictly increasing, continuously differentiable real-valued function with \(b(0)=0 \) and \(b'(t) >0\) for each \(t \in [0,T]\).
As explained in [11, pp. 18–20], Y induces a probability measure \(\mu \) on the measurable space \((\mathbb {R}^D,\mathcal {B}^D )\) where \(\mathbb {R}^D\) is the space of all real-valued functions x(t), \(t \in D\), and \(\mathcal {B}^D\) is the smallest \(\sigma \)-algebra of subsets of \(\mathbb {R}^D\) with respect to which all the coordinate evaluation maps \(e_t (x)=x(t)\) defined on \({\mathbb R}^D\) are measurable. The triple \(({\mathbb R}^D,\mathcal {B}^D,\mu )\) is a probability measure space. This measure space is called the function space induced by the generalized Brownian motion process Y determined by \(a(\cdot )\) and \(b(\cdot )\).
We note that the generalized Brownian motion process Y determined by \(a(\cdot )\) and \(b(\cdot )\) is a Gaussian process with mean function a(t) and covariance function \(r(s,t)= \min \{b(s),b(t)\}\). By [11, Theorem 14.2], the probability measure \(\mu \) induced by Y, taking a separable version, is supported by \( C_{a,b}[0,T]\) (which is equivalent to the Banach space of continuous functions x on [0, T] with \(x(0)=0\) under the sup norm). Hence, \((C_{a,b}[0,T],{\mathcal {B}}(C_{a,b}[0,T]),\mu )\) is the function space induced by Y where \({\mathcal {B}}(C_{a,b}[0,T])\) is the Borel \(\sigma \)-algebra of \(C_{a,b}[0,T]\). We then complete this function space to obtain \((C_{a,b}[0,T],{\mathcal {W}}(C_{a,b}[0,T]),\mu )\) where \({\mathcal {W}}(C_{a,b}[0,T])\) is the set of all \(\mu \)-Carathéodory measurable subsets of \(C_{a,b}[0,T]\).
We note that the coordinate process defined by \(e_t(x)=x(t)\) on \(C_{a,b}[0,T]\times [0,T]\) is also the generalized Brownian motion process determined by a(t) and b(t). For more detailed studies about this function space \(C_{a,b}[0,T]\), see [3, 6, 10].
Next, we state the definition of the analytic operator-valued generalized Feynman integral.
Definition 2.1
Let \(\mathbb {C}\) be the set of complex numbers, let \({\mathbb C}_{+}=\{\lambda \in \mathbb {C}{:}{} \textit{Re}(\lambda )>0\}\) and let \(\tilde{\mathbb C}_{+}=\{\lambda \in \mathbb {C}{:}{} \textit{Re}(\lambda )\ge 0,\lambda \not =0\}\). Also, let C[0, T] denote the space of real-valued continuous functions x on [0, T], and given a real number \(\alpha \), let \(\nu _{\alpha }\) be the measure on \(\mathcal {B}(\mathbb {R})\) such that \(\hbox {d}\nu _{\alpha } =\exp \{\alpha \eta ^2\}\hbox {d}\eta \). Next let F be a \(\mathbb {C}\)-valued functional on C[0, T]. For each \(\lambda >0\), \(\psi \in L^2(\mathbb {R},\nu _{\alpha })\) and \(\xi \in \mathbb {R}\), assume that the functional \(F(\lambda ^{-1/2}x+\xi )\psi (\lambda ^{-1/2}x(T)+\xi )\) is \(\mu \)-integrable with respect to x on \(C_{a,b}[0,T]\), and let
If \(I_{\lambda }(F)\psi \) is in \(L^{2}(\mathbb {R},\nu _{-\alpha })\) as a function of \(\xi \) and if the correspondence \(\psi \rightarrow I_{\lambda }(F)\psi \) gives an element of \(\mathcal {L}\equiv \mathcal {L}(L^{2}(\mathbb {R},\nu _{\alpha }),L^{2}(\mathbb {R},\nu _{-\alpha }))\), the space of continuous linear operators from \(L^2(\mathbb {R},\nu _{\alpha })\) to \(L^2(\mathbb {R},\nu _{-\alpha })\), we say that the operator-valued function space integral \(I_{\lambda }(F)\) exists. Next, suppose that there exists an \(\mathcal {L}\)-valued function which is analytic in \(\mathbb {C}_{+}\) and agrees with \(I_{\lambda }(F)\) on \((0,\infty )\), then this \(\mathcal {L}\)-valued function is denoted by \(I_{\lambda }^{\mathrm{an}}(F)\) and is called the analytic operator-valued function space integral of F associated with \(\lambda \). Finally, suppose that there exists an operator \(J_{q}^{\mathrm{an}}\) in \(\mathcal {L}(L^2(\mathbb {R},\nu _{\alpha }),L^2(\mathbb {R},\nu _{-\alpha }))\) for some \(\alpha >0\) such that
as \(\lambda \rightarrow -iq\) through \({\mathbb C}_{+}\), then \(J_{q}^{\mathrm{an}}(F)\) is called the analytic operator-valued generalized Feynman integral of F with parameter q.
3 An Analytic Operator-Valued Function Space Integral
Throughout the rest of this paper, we consider functionals of the form
where f is an analytic function on \(\mathbb C\) and \(\theta \) is an appropriate \(\mathbb C\)-valued function on \([0,T]\times \mathbb {R}\). F(x) is a very important functional in quantum mechanics. We then establish the existence of the analytic operator-valued function space integral for functionals F of the form (3.1).
Let \(\mathcal {M}(0,T)\) denote the space of complex Borel measures \(\eta \) on the open interval (0, T). Then \(\eta \in \mathcal {M} (0,T) \) has a unique decomposition \(\eta = \beta + \beta _d \) into its continuous part \(\beta \) and its discrete part \(\beta _d\) [9]. Let \(\delta _{\tau }\) denote the Dirac measure at \(\tau \in (0,T)\). For convenience, we let
Throughout the rest of this paper, we use the following notations:
(1) For \(\lambda \in \tilde{\mathbb C}_+\) and \(\psi \in L^2(\mathbb {R},\nu _{\alpha })\), let
where K and L are real numbers with \(K>0\). Then \(C_{(\lambda ,K,L)}\) is in \(\mathcal {L}(L^2(\mathbb {R},\nu _{\alpha }), L^2(\mathbb {R},\nu _{-\alpha }) )\).
(2) For each \(s\in (0,T)\), let \(\theta (s)\) denote the operator of multiplication from \(L^2(\mathbb {R},\nu _{-\alpha })\) to \(L^2(\mathbb {R},\nu _{\alpha })\) given by
(3) Given a positive integer \(l_1\), let
and let
Also, for \((s_1,\ldots , s_{l_1})\in \Delta _{l_1 ; j}(T)\) and a positive integer \(l_2\), let
Finally, for \((s_1,\ldots ,s_{l_1})\in \Delta _{l_1}(T)\), let
For example, we see that for \(s_1\in \Delta _{1;1}(T)=\{s_1| 0< s_1<\tau <T\}\),
and for \((s_1,s_2)\in \Delta _2(T)\),
Hence using Eqs. (3.3)–(3.6), we observe that for \(\psi \in L^2(\mathbb {R},\nu _{\alpha })\),
and
where \(s_0=0\), \(a(s_0)=0\), \(u_0 = \xi \), \(s_2 = \tau \) and \(s_3 = T\).
Also we will use the following conventions: for all positive integer l and \(\lambda \in \tilde{\mathbb C}_{+}\), let
for some \(M_j>0\), \(j=1,\ldots ,{l_1}\). Furthermore, in order to ensure that analytic operator-valued generalized Feynman integral exists, we will assume that \(B_j^l (s_j;|\lambda |)\), \(a(\cdot )\) and \(b(\cdot )\) satisfy the following conditions: for \(j=1,\ldots ,l_1\) and \(s_{l_1+1}=T\),
-
(1)
\(\int _{0}^{T} B_j^l \left( s_j;\left| \lambda \right| \right) \hbox {d}|\eta |(s) < \infty \)
-
(2)
\(\frac{1}{b(s_j) - b(s_{j-1})} \le L_{jn} \)
-
(3)
\(\left| a'(s_j^*)\right| \le \left| b'(s_j^*)\right| M_{jn} \)
for \(s_j^* \in (s_{j-1},s_j)\) and some positive real numbers \(L_{jn}\) and \(M_{jn}\).
The next lemma plays a key role in the proof of Theorem 3.2.
Lemma 3.1
Let \(\mathcal {L}^{ \lambda }_{l_1;j}\) be given by Eq. (3.5). Then for all \(l_2\in \mathbb {N}\), \(\xi \in \mathbb {R}\), \(\lambda \in \tilde{\mathbb {C}}_+\) and \(\psi \in L^2(\mathbb {R},\nu _{\alpha })\),
for some \(\alpha >0\).
Proof
Using Eq. (3.3)–(3.5), we have that for all \(l\in \mathbb {N}\), \(\xi \in \mathbb {R}\) and \(\lambda \in \tilde{\mathbb {C}}_+\)
which completes the proof of Lemma 3.1. \(\square \)
In our next theorem, we establish the existence of the analytic operator-valued function space integral for the functional F given by (3.1) with \(f(z)=z^n\).
Theorem 3.2
Let \(\theta \) be a Borel measurable function on \([0,T]\times \mathbb {R}\). For \(n=1,2,\ldots \), let
Let \(\eta \) be given by (3.2). Then for all \( \lambda \in \mathbb {C}_+\) and \(\psi \in L^2(\mathbb {R},\nu _{\alpha })\), the analytic operator-valued function space integral of \(F_n\), \(I_{ \lambda }^{\mathrm{an}}(F_n)\), exists and is given by the formula
where \(\beta (s_0)=0\), \(s_{l_1 +1} = T\) and \(\Delta _{0;j}(T)\) is an empty set.
Proof
Using Eq. (3.1) with \(f(z)=z^n\), (3.3), (3.4) and the Fubini theorem, we first obtain that for all \( \lambda >0\)
Next we will show that the existence of analytic operator-valued function space integral \(I_{\lambda }^{\mathrm{an}}(F_n)\) exists. Using Eq. (3.8), we obtain that for all \(\lambda \in \mathbb {C}_+\)
Therefore, the analytic operator-valued function space integral \(I_{\lambda }^{\mathrm{an}}(F_n)\) exists and is given by Eq. (3.10).
Now we will show that \(I_{\lambda }^{\mathrm{an}}(F_n)\) is an element of \(\mathcal {L}( L^2(\mathbb {R}, \nu _{\alpha }), L^2(\mathbb {R}, \nu _{-\alpha }))\). Using Eqs. (3.10) and (3.11), it follows that
Hence, we obtain that for all \(\lambda \in \mathbb {C}_+\),
Thus, the theorem is proved. \(\square \)
Let \(f(z) = \sum ^{\infty }_{n=1} \beta _n z^n\) be an analytic function on \(\mathbb C\) such that
for all \(\lambda \in \tilde{\mathbb C}_{+}\), where
for all positive integers n and k. Let
for \(x\in C_{a,b}[0,T]\).
Our aim in this section is to establish the existence of the analytic operator-valued function space integral for the functionals F given by (3.15).
Theorem 3.3
Let F be given by Eq. (3.15). Then for all \(\lambda \in \mathbb {C}_+\) and \(\psi \in L^2(\mathbb {R},\nu _{\alpha })\), the analytic operator-valued function space integral of F, \(I_{\lambda }^{\mathrm{an}}(F)\), exists and is given by the formula
where \(I_{\lambda }^{\mathrm{an}}(F_n)\) is given by Eq. (3.10). Furthermore, \(I_{\lambda }^{\mathrm{an}}(F)\) is an element of \( \mathcal {L} ( L^2(\mathbb {R},\nu _{\alpha }), L^2(\mathbb {R},\nu _{-\alpha }) )\).
Proof
Since \(F(x) = \sum _{n=1}^{\infty }\beta _n F_n(x)\), using (3.11) and (3.12) we have
and
where \(\Psi ^1_{n}(|\lambda |)\) is given by Eq. (3.14) with \(k=1\). Next using the condition (3.13), the analytic operator-valued function space integral \(I_{\lambda }^{\mathrm{an}}(F) \) exists and \(I_{\lambda }^{\mathrm{an}}(F)\) is an element of \(\mathcal {L}(L^2(\mathbb {R},\nu _{\alpha }),L^2(\mathbb {R},\nu _{-\alpha }))\). \(\square \)
4 An Analytic Operator-Valued Generalized Feynman Integral
In Sect. 3, we established the existence of the analytic operator-valued function space integral for the functionals F given by Eq. (3.15). In this section, we establish the existence of the analytic operator-valued generalized Feynman integral for the functionals F. To do this, in Theorem 4.1, we first obtain the analytic operator-valued generalized Feynman integral for the functionals \(F_n\) given by (3.9).
Theorem 4.1
Let \(F_n\) be given by Eq. (3.9). Then for all \(q\in \mathbb {R} {\setminus }\{0\}\), the analytic operator-valued generalized Feynman integral of \(F_n\), \(J_{q}^{\mathrm{an}}(F_n)\), exists and is given by the formula
where \(\beta (s_0)=0\), \(s_{l_1 +1} = T\) and \(\Delta _{0;j}(T)\) is an empty set.
Proof
In order to establish Eq. (4.1), it suffices to show that
But, for all \(\lambda \in \mathbb {C}_+\), we have
Using a similar method as those used in (3.12), we also see that \(|(I_{\lambda }^{\mathrm{an}}(F_n))(\psi )|^2\) and \( |(J_{q}^{\mathrm{an}}(F_n))(\psi )|^2\) are in \(L^1(\mathbb {R},\nu _{-\alpha })\). Hence, the second expression in Eq. (4.2) is in \(L^1(\mathbb {R},\nu _{-\alpha })\). Thus, using the dominated convergence theorem, we obtain the desired result. \(\square \)
The next theorem is one of the main results in this paper.
Theorem 4.2
Let F be given by Eq. (3.15). Then for all \(q\in \mathbb {R}{\setminus }\{0\}\), the analytic operator-valued generalized Feynman integral of F, \(J_{q}^{\mathrm{an}}(F)\), exists and is given by the formula
where \(J_{q}^{\mathrm{an}}(F_n)\) is given by Eq. (4.1). Furthermore, \(J_{q}^{\mathrm{an}}(F)\) is an element of \(\mathcal {L}(L^2(\mathbb {R},\nu _{\alpha }), L^2(\mathbb {R},\nu _{-\alpha }) )\).
Proof
Using (3.11) and (3.12) with \(\lambda \) replaced with \(-iq\), we obtain
and
where \(\Psi ^1_n(|-iq|)\) is given by Eq. (3.14) with \(k=1\). Next using the condition (3.13), we conclude that the analytic operator-valued generalized Feynman integral \(J_{q}^{\mathrm{an}}(F) \) exists and \(J_{q}^{\mathrm{an}}(F)\) is an element of \(\mathcal {L}(L^2(\mathbb {R},\nu _{\alpha }),L^2(\mathbb {R},\nu _{-\alpha }))\). \(\square \)
The next two lemmas play key roles in the proof of Theorem 4.5.
Lemma 4.3
For each \(k=1,2,\ldots \), let \(F_n^{(k)}\) be given by (3.9) with \(\theta \) replaced with \(\theta ^{(k)}\). Then for all \(q\in \mathbb {R}{\setminus }\{0\}\), the analytic operator-valued generalized Feynman integral of \(F_n^{(k)}\), \(J_{q}^{\mathrm{an}}(F_n^{(k)})\), exists and is given by the formula
where \(\mathcal {L}^{-iq}_{l_1;j;k}\) is given by the right-hand side of Eq. (3.5) with \(\theta \) replaced by \(\theta ^{(k)}\). Furthermore, we have
where \(F^{(k)}:C_{a,b}[0,T]\rightarrow \mathbb C\) is given by
for each \(k=1,2,\ldots .\)
Proof
The proof is straightforward by replacing \(\theta \) with \(\theta ^{(k)}\) in Theorem 4.1. \(\square \)
Lemma 4.4
Let \(F_n^{(k)}\) be as in Lemma 4.3. Then for all \(q\in \mathbb {R}{\setminus }\{0\}\) and \(\psi \in L^2(\mathbb {R},\nu _{\alpha })\),
Proof
To establish Eq. (4.6) it will suffice to show that
for all \(\psi \in L^2(\mathbb {R},\nu _{\alpha })\). But using similar methods as those used in (3.11), it follows that for each \(n\in \mathbb N\),
where \(B^{(k)}\) is given by Eq. (3.7) with \(\theta \) replaced with \(\theta ^{(k)}\). Also, the last expression of (4.7) is in \(L^2(\mathbb {R},\nu _{-\alpha })\) and it dominates the sequence of functions \(|(J_{q}^{\mathrm{an}}(F_n^{(k)}) \psi )(\xi )-(J_{q}^{\mathrm{an}}(F_n) \psi )(\xi ) |^2\). Hence using the dominated convergence theorem, we obtain the desired result. Furthermore, using similar methods as those used in (3.12) we have
and
where \(\Psi ^k_n(|-iq|)\) is given by Eq. (3.14). \(\square \)
We are now ready to establish our main result, namely the stability theorem for the analytic operator-valued generalized Feynman integral.
Theorem 4.5
Let \(\{\theta ^{(k)}\}\) be a sequence of complex-valued functions such that \(\theta ^{(k)}(s,u)\rightarrow \theta (s,u)\), as \(k\rightarrow \infty \), for \(\eta \times m_{L}\)-a.e. (s, u). For \(k=1,2,\ldots \), let the functional \(F^{(k)}\) on \(C_{a,b}[0,T]\) be given by Eq. (4.5). Then for all \(q\in \mathbb {R}{\setminus }\{0\}\) and \(\psi \in L^2(\mathbb {R},\nu _{\alpha })\),
where \(J_{q}^{\mathrm{an}}(F^{(k)})\) is given by Eq. (4.4).
Proof
Using Eqs. (4.3), (4.4) and (4.6) we have that
is in \(L^2(\mathbb {R},\nu _{-\alpha })\). Step \(\mathrm (I)\) follows from Lemma 4.3. From Eqs. (3.13) and (4.8), we have
Also, by using Eqs. (4.6) and (4.8), we can show that \(J_{q}^{\mathrm{an}} (F_n^{(k)}) \psi \rightarrow J_{q}^{\mathrm{an}} (F_n ) \psi \) in \(L^2(\mathbb {R},\nu _{-\alpha })\) as \(k\rightarrow \infty \), and hence, \(J_{q}^{\mathrm{an}} (F_n ) \psi \) exists. Hence, Step \(\mathrm (II)\) now follows. From Lemma 4.4, we obtain Step \(\mathrm (III)\). Step \(\mathrm (IV)\) then follows from Theorem 4.2. \(\square \)
References
Cameron, R.H., Storvick, D.A.: An operator valued function space integral and a related integral equation. J. Math. Mech. 18, 517–552 (1968)
Chang, K.S., Ko, J.W., Ryu, K.S.: Stability theorems for the operator-valued Feynman integral: the \(\cal{L}(L_1(\mathbb{R}), C_0(\mathbb{R}))\) theory. J. Korean Math. Soc. 35, 999–1018 (1998)
Chang, S.J., Choi, J.G., Skoug, D.: Integration by parts formulas involving generalized Fourier–Feynman transforms on function space. Trans. Am. Math. Soc. 355, 2925–2948 (2003)
Chang, S.J., Chung, D.M.: Conditional function space integrals with applications. Rocky Mt. J. Math. 26, 37–62 (1996)
Chang, S.J., Lee, I.Y.: Analytic operator-valued generalized Feynman integrals on function space. J. Chungcheong Math. Soc. 23, 37–48 (2010)
Chang, S.J., Skoug, D.: Generalized Fourier–Feynman transforms and a first variation on function space. Integral Transforms Special Funct. 14, 375–393 (2003)
Johnson, G.W.: A bounded convergence theorem for the Feynman integral. J. Math. Phys. 25, 1323–1326 (1984)
Johnson, G.W., Lapidus, M.L.: Generalized Dyson series, generalized Feynman diagrams, the Feynman integral and Feynman’s operational calculus. Mem. Am. Math. Soc. 62, 1–78 (1986)
Reed, M., Simon, B.: Methods of Modern Mathematical Physics, vol. I, Rev. and enlarged edn. Academic Press, New York (1980)
Yeh, J.: Singularity of Gaussian measures on function spaces induced by Brownian motion processes with non-stationary increments. Ill. J. Math. 15, 37–46 (1971)
Yeh, J.: Stochastic Processes and the Wiener Integral. Marcel Dekker Inc., New York (1973)
Acknowledgements
The authors thank the referees for their helpful suggestions which led to the present version of this paper.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Rosihan M. Ali.
Rights and permissions
About this article
Cite this article
Chang, S.J., Choi, J.G. & Lee, I.Y. An Analytic Operator-Valued Generalized Feynman Integral on Function Space. Bull. Malays. Math. Sci. Soc. 42, 521–534 (2019). https://doi.org/10.1007/s40840-017-0498-4
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40840-017-0498-4
Keywords
- Analytic operator-valued function space integral
- Analytic operator-valued generalized Feynman integral
- Stability theorem