Abstract
A time-inhomogeneous Feller-type diffusion process with linear infinitesimal drift \(\alpha (t)x+\beta (t)\) and linear infinitesimal variance 2r(t)x is considered. For this process, the transition density in the presence of an absorbing boundary in the zero-state and the first-passage time density through the zero-state are obtained. Special attention is dedicated to the proportional case, in which the immigration intensity function \(\beta (t)\) and the noise intensity function r(t) are connected via the relation \(\beta (t)=\xi \,r(t)\), with \(0\le \xi <1\). Various numerical computations are performed to illustrate the effect of the parameters on the first-passage time density, by assuming that \(\alpha (t)\), \(\beta (t)\) or both of these functions exhibit some kind of periodicity.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction and Background
One-dimensional time-inhomogeneous diffusion processes play a relevant role in different application fields, including physics, biology, neuroscience, finance and others (cf., for instance, Giorno and Nobile [1, 2], Albano and Giorno [3], Ghost and Prajneshu [4], Buonocore et al [5], Gutiérrez et al [6], Di Crescenzo et al [7], Román-Román et al. [8], Molini et al. [9], Gan and Waxman [10], Abundo [11]). In this paper, we consider a time-inhomogeneous Feller-type diffusion process, characterized by linear infinitesimal drift and linear infinitesimal variance vanishing in the zero-state (lower boundary of the process). We assume that the zero-state represents an absorbing boundary for the process.
Let \(\{X(t),t\ge t_0\}\), \(t_0\ge 0\), be a time-inhomogeneous Feller-type diffusion process with the following infinitesimal drift and infinitesimal variance
defined in the state-space \([0,+\infty )\), with \(\alpha (t)\in {\mathbb {R}}\), \(\beta (t)\in {\mathbb {R}}\), \(r(t)>0\) continuous functions for all \(t\ge t_0\).
The time-homogeneous Feller diffusion process, in which \(\alpha (t)=\alpha \), \(\beta (t)=\beta \) and \(r(t)=r\) for all \(t\ge 0\), is taken in account in Feller [12], where is shown that boundary \(x=0\) changes its character depending on whether \(\beta \le 0\) (exit), \(0< \beta < r\) (regular), \(\beta \ge r\) (entrance). Furthermore, as proved in Feller [13], if one knows the nature of the end points of the state-space one can decide what kind of boundary condition has to be associated with the Fokker-Planck and Kolmogorov diffusion equations to determine the transition pdf of the process. A review showing the relevance of the Feller’s work on boundary classification of one-dimensional diffusion processes is provided in Peskir [14]. By following this approach, for the time-homogeneous Feller diffusion process, the transition pdf in the presence of an absorption condition or a zero-flux condition in the zero-state is explicitly obtained in Karlin and Taylor [15] and in Giorno et al. [16]. Furthermore, a class of Kolmogorov diffusion equations that can be transformed into a Kolmogorov equation for a time-homogeneous Feller process is considered in Capocelli and Ricciardi [17].
Feller diffusion process is widely used in mathematical biology to model the growth of a population (cf., Lavigne and Roques [18], Masoliver [19], Ricciardi et al. [20]), in queueing systems to describe the number of customers in a queue (cf., Di Crescenzo and Nobile [21]), in neurobiology to analyze the input-output behavior of single neurons (see, for instance, Ditlevsen and Lánský [22], Lánský et al. [23], Nobile and Pirozzi [24], Giorno et al. [25, 26], Buonocore et al. [27]), in mathematical finance to model asset prices, market indices, interest rates and stochastic volatility (see, Tian and Zhang [28], Cox et al. [29], Linetsky [30], Göing-Jaeshke and Yor [31]).
Sometimes, the Feller-type diffusion process X(t) is obtained as a continuous approximation of a time-inhomogeneous discrete Markov processes (see, for instance, Di Crescenzo and Nobile [21], Giorno et al. [25]). Indeed, in population dynamics the Feller-type diffusion process arises as a continuous approximation of a birth-death process with immigration (cf. Giorno and Nobile [32] and references therein). In these cases \(\alpha (t)\), related to the growth intensity function, is positive (negative) when the birth intensity function is greater (less) than the death intensity function, whereas \(\alpha (t)=0\) if the birth intensity function is equal to the death intensity function. Since \(\alpha (t)\) is a time dependent function, it can be positive, negative or zero at different time instants. Instead, \(\beta (t)\) is related to the immigration intensity function. In particular, \(\beta (t)>0\) indicates the presence of immigrations and a zero-flux condition or an absorbing condition can be imposed in the zero-state of the diffusion process.
For a full characterization of the time-inhomogeneous Feller-type diffusion process X(t), the behavior at the boundary 0 must be specified. In this paper, we assume that the zero-state is an absorbing boundary, so that the process X(t) terminates when the boundary is reached. We suppose that \(\alpha (t)\in {\mathbb {R}}\), \(\beta (t)\in {\mathbb {R}}\), \(r(t)>0\), \(\beta (t)\le \xi \, r(t)\), with \(0\le \xi <1\), for all \(t\ge t_0\).
We denote by
the transition probability density function (pdf) of X(t) in the presence of an absorbing boundary in the zero-state. As shown in Kolmogorov [33] and Dynkin [34], the pdf \(f_a(x,t|x_0,t_0)\) satisfies the Kolmogorov equation
with \(A_1(x_0,t_0)\) and \(A_2(x_0,t_0)\) given in (1), to solve imposing the initial delta condition
and the absorbing boundary condition in the zero-state:
Furthermore, let
be the random variable describing the first-passage time (FPT) through the zero-state starting from \(X(t_0)=x_0>0\); we denote by
We note that the FPT density \(g(0,t|x_0,t_0)\) is not affected by the boundary condition on the zero-state, provided that it is attainable.
The problem of determining FPT densities for the Feller-type diffusion process arises in a variety of fields, including neurobiology, population dynamics, queueing systems and mathematical finance (cf., for instance, Linetsky [30], Masoliver and Perelló [35], Buonocore et al. [36], D’Onofrio et al. [37], Giorno et al. [38, 39], Albano e Giorno [40], Di Nardo and D’Onofrio [41]). For instance, in population dynamics \(g(0,t|x_0,t_0)\) describes the extinction density, whereas in queueing systems represents the busy period density. Lavigne and Roques in [18] focus on the distribution of the extinction times of a population whose size is described by a time-inhomogeneous Feller-type diffusion process with infinitesimal drift \(A_1(x,t)=\alpha (t)\,x\) and infinitesimal variance \(A_2(x,t)=\sigma ^2\,x\), where \(\alpha (t)\) is a continuous function and \(\sigma ^2\) is a positive constant.
The functions (2) and (7) are intimately related; indeed, one has:
Relation (8) shows that the determination of \(g(0,t|x_0,t_0)\) requires the explicit evaluation of the transition pdf \(f_a(x,t|x_0,t_0)\) in the presence of an absorbing boundary at the zero-state.
1.1 Plain of the Paper
The paper is organized in five sections and seven appendices in which the proofs of the main results are reported. In Sect. 2, for the time-inhomogeneous Feller-type diffusion process X(t), with infinitesimal moments (1), we give some preliminary results concerning the Laplace transform (according to \(x_0\)) of the transition pdf \(f_a(x,t|x_0,t_0)\) in the presence of an absorbing boundary in the zero-state. The proportional case, in which the immigration intensity function \(\beta (t)\) and the noise intensity function r(t) are related as \(\beta (t)=\xi \,r(t)\), with \(0\le \xi <1\), is also analyzed. In Sect. 3, the transition pdf \(f_a(x,t|x_0,t_0)\) is obtained for the process (1) in the general case, by distinguishing the case \(x=0\) (Sect. 3.1) and \(x>0\) (Sect. 3.2). In Sect. 4, we focus on the FPT of X(t) through the zero-state for the general case and we determine the expression of the FPT pdf \(g(0,t|x_0,t_0)\). In Sects. 3 and 4, we also show as the results of the proportional case can be derived from the general case. In Sect. 5, various numerical computations are performed making use of MATHEMATICA to illustrate the effect of periodic intensity functions on the FPT pdf \(g(0,t|x_0,t_0)\). Specifically, we assume that the growth intensity function \(\alpha (t)\), the immigration intensity function \(\beta (t)\) or both these functions exhibit some kind of periodicity. The FPT mean \(t_1(0,t|x_0,t_0)\) and the coefficient of variation \(\mathrm{CV}(0|x_0,t_0)=\sqrt{ \mathrm{Var}(0|x_0,t_0)} / t_1(0|x_0,t_0)\) are also analyzed.
2 Preliminary Results
In this section, we determine the Laplace transform (according to \(x_0\)) of the transition pdf \(f_a(x,t|x_0,t_0)\) in the general case. Furthermore, the explicit expressions of the transition pdf and of the FPT density through the zero-state are obtained in the proportional case.
2.1 Laplace Transform
For \(t\ge t_0\) and \(x\ge 0\), we consider the Laplace transform:
We determine \(Z_a(x,t|s,t_0)\) so that, by taking its inverse Laplace transform, we obtain \(f_a(x,t|x_0,t_0)\). Multiplying both sides of (3) by \(e^{-sx_0}\), integrating with respect to \(x_0\) over the interval \([0,+\infty )\) and making use of the boundary condition (5), we have the following partial differential equation
to solve with the initial condition
derived from (9) by using the initial condition (4).
Proposition 1
We assume that \(\alpha (t)\in {\mathbb {R}}\), \(\beta (t)\in {\mathbb {R}}\), \(r(t)>0\), \(\beta (t)\le \xi \, r(t)\), with \(0\le \xi <1\). For \(t\ge t_0\), we have:
where
Proof
The proof is given in Appendix A. \(\square \)
2.2 Proportional Case
For all \(t\ge 0\), we suppose that the continuous functions \(\beta (t)\) and r(t) are proportional, i.e.
In the absence of immigration, i.e. when \(\beta (t)=0\) for all \(t\ge 0\), one has \(\xi =0\).
Proposition 2
Under the assumption (14), for \(t\ge t_0\) one has:
Furthermore, the transition pdf of X(t) in the presence of an absorbing boundary in the zero-state is:
with \(A(t|t_0)\) and \(R(t|t_0)\) defined in (13) and where
denotes the modified Bessel function of the first kind.
Proof
The proof is given in Appendix B. \(\square \)
Note that, the first of (16) follows by taking the limit as \(x\downarrow 0\) in the second, recalling that for fixed \(\nu \) and for \(z\rightarrow 0\) one has (cf. Abramowitz and Stegun [42], p. 375, no 9.6.7):
If (14) holds, for \(t\ge t_0\), \(x>0\) and \(x_0>0\) from (16) it follows:
Proposition 3
Under the assumption (14), for \(t\ge t_0\) and \(x_0>0\) one has:
with \(R(t|t_0)\) given in (13) and where
denotes the incomplete gamma function.
Proof
Recalling (16) and using the transformation \(y=x\,e^{-A(t|t_0)}/R(t|t_0)\) in the integral, one obtains:
Since (cf. Erdèlyi et al. [43], p. 197, no. 19)
Eq. (19) follows from (21). \(\square \)
Proposition 4
Under the assumption (14), for \(t\ge t_0\) and \(x_0>0\) the FPT pdf through the zero-state of X(t) is:
with \(R(t|t_0)\) given in (13). Furthermore, for \(0\le \xi <1\) the ultimate FPT probability is:
Proof
By virtue of (8), making use of (19), for \(0\le \xi <1\) one has:
from which (22) follows. Furthermore, taking the limit as \(t\rightarrow +\infty \) in (8), it results
so that, recalling (19), one is lead to (23). \(\square \)
Two interesting cases occur when \(\xi =0\) and \(\xi =1/2\).
Indeed, by setting \(\xi =0\) in (14), one considers the time-inhomogeneous Feller-type diffusion process (1) with \(\beta (t)=0\). In the context of population dynamics, this case describes the absence of the immigration and it is of interest to determine for which choices of \(\alpha (t)\) and r(t) the population is doomed to extinction as the time increases. Recalling that \(\gamma (1,z)=1-e^{-z}\), for \(t\ge t_0\), \(x_0>0\) and \(\xi =0\), from (19) one has
and from (22) one obtains:
Instead, when \(\xi =1/2\) in (14), the FPT pdf (22) identifies with the FPT pdf through the zero-state for special time-inhomogeneous Wiener or Ornstein-Uhlenbeck diffusion processes (see, for instance, Giorno and Nobile [1]). Specifically, if \(\xi =1/2\) and \(\alpha (t)=0\), the density (22) identifies with the FPT pdf \(g_{W}(0,t|\sqrt{x_0},t_0)\) of a time-inhomogeneous Wiener process, with state-space in \({\mathbb {R}}\), having infinitesimal drift \(B_1(x,t)=0\) and infinitesimal variance \(B_2(t)=r(t)/2\); instead, if \(\xi =1/2\) and \(\alpha (t)\ne 0\), the density (22) identifies with the FPT pdf \(g_{OU}(0,t|\sqrt{x_0},t_0)\) of a time-inhomogeneous Ornstein-Uhlenbeck process, with state-space in \({\mathbb {R}}\), having infinitesimal drift \(C_1(x,t)=[\alpha (t)/2]\,x\) and infinitesimal variance \(C_2(t)=r(t)/2\).
Under the assumption (14), if \(\lim _{t\rightarrow +\infty } R(t|t_0)=+\infty \), it is meaningful to evaluate the FPT moments through the zero-state starting from \(X(t_0)=x_0>0\):
Indeed, if \(\lim _{t\rightarrow +\infty } R(t|t_0)=+\infty \), from Proposition 4 one has \(P\{T(x_0,t_0)<+\infty \}=1\) and, making use of (8) and (19), for \(0\le \xi <1\) one has:
We finally note that, for the time-homogeneous Feller process, in which \(\alpha (t)=\alpha \), \(\beta (t)=\xi \,r\), \(r(t)=r\), with \(\alpha \in {\mathbb {R}}\), \(r>0\) and \(0\le \xi <1\), the pdf \(f_a(x,t|x_0,t_0)\) and the FPT pdf \(g(0,t|x_0,t_0)\) can be easily obtained from (16) and (22) by setting
3 General Case
We assume that \(\alpha (t)\), \(\beta (t)\) and r(t) are continuous functions such that \(\alpha (t)\in {\mathbb {R}}\), \(\beta (t)\in {\mathbb {R}}\), \(r(t)>0\), \(\beta (t)\le \xi \, r(t)\), with \(0\le \xi <1\). From (12), for \(t\ge t_0\) we have
where
with \(A(t|t_0)\) and \(R(t|t_0)\) given in (13). We note that \(V_a(x,t|s,t_0)\) does not dependent upon \(\beta (t)\). Therefore, to obtain the transition pdf \(f_a(x,t|x_0,t_0)\) for X(t) with infinitesimal moments (1), we proceed as follows:
-
(1)
we determine the transition pdf \(f_a(0,t|x_0,t_0)\) for \(x_0>0\) and \(t\ge t_0\) by taking the inverse Laplace transform of \(Z_a(0,t|s,t_0)\);
-
(2)
we find the inverse Laplace transform \(v_a(x,t|x_0,t_0)\) of (29) and we calculate the transition pdf \(f_a(x,t|x_0,t_0)\) as a convolution, according to \(x_0\), between \(f_a(0,t|x_0,t_0)\) and the function \(v_a(x,t|x_0,t_0)\) for \(x>0\), \(x_0>0\) and \(t\ge t_0\).
3.1 General Case: \(x=0\)
In this section, we obtain the transition pdf in the presence of an absorbing boundary in the zero-state when the process X(t) reaches \(x=0\) at time \(t\ge t_0\). By setting \(x=0\) in (12), for \(t\ge t_0\) we obtain:
with \(A(t|t_0)\) and \(R(t|t_0)\) defined in (13).
In the sequel, we denote by \(B_n(d_1,d_2,\dots ,d_n)\) the complete Bell polynomials, recursively defined as follows:
with
Proposition 5
Under the assumption of Proposition 1, for \(t\ge t_0\) and \(x_0>0\) the transition pdf of the time-inhomogeneous Feller-type diffusion process X(t) with an absorbing boundary in the zero-state is
where
with \(A(t|t_0)\) and \(R(t|t_0)\) defined in (13), \(B_n(d_1,d_2,\dots ,d_n)\) given in (31) and in (32), and
denoting the Laguerre polynomials.
Proof
The proof is given in Appendix C. \(\square \)
Remark 1
(Proportional case) We assume that (14) holds. We prove that the first of (16) follows from (33).
Indeed, from (31) and (32) one has
where \((\xi )_n\) denotes the Pochhammer symbol defined as \((\xi )_0=1\) and \((\xi )_n=\xi \,(\xi +1)\cdots (\xi +n-1)\) for \(n=1,2,\ldots \) Recalling that the series of Laguerre polynomials satisfies the following identity (cf. Erdèlyi et al [44], p. 213, no. 16):
under the assumption (14), from (34) one has:
Hence, if (14) holds, Eq. (33) identifies with the first of (16). \(\Diamond \)
3.2 General Case: \(x>0\)
In this section, we obtain the transition pdf \(f_a(x,t|x_0,t_0)\) for \(x>0\) and \(t\ge t_0\). From (29), for \(t\ge t_0\), \(x>0\) and \({\text {Re}} s>0\) we have \(V_a(x,t|s,t_0)\ge 0\) and
We show that \(V_a(x,t|s,t_0)\) is the Laplace transform of a function \(v_a(x,t|x_0,t_0)\), i.e.
Proposition 6
Under the assumption of Proposition 1, for \(x_0>0\) and \(t\ge t_0\), one has:
with \(A(t|t_0)\) and \(R(t|t_0)\) defined in (13), whereas \(\delta (x)\) denotes the delta Dirac function and \(I_\nu (z)\) represents the Bessel function modified of first kind.
Proof
The proof is given in Appendix D. \(\square \)
The function \(v_a(x,t|x_0,t_0)\) in (40) is the sum of two terms. The second term in (40) identifies with \(x\,f_a(x,t|x_0,t_0)/x_0\), where \(f_a(x,t|x_0,t_0)\) is given in (16) for \(x>0\) and \(\xi =0\) (absence of immigration). Since,
from (40) it follows that
For \(x>0\), the transition pdf \(f_a(x,t|x_0,t_0)\) can be obtained via a convolution, according to \(x_0\), between the pdf \(f_a(0,t|x_0,t_0)\) and the function \(v_a(x,t|x_0,t_0)\), determined in Propositions 5 and 6, respectively:
Proposition 7
Under the assumption of Proposition 1, for \(t\ge t_0\), \(x>0\) and \(x_0>0\) one has:
with \(A(t|t_0)\) and \(R(t|t_0)\) given in (13) and \(\varPsi (t|z,t_0)\) defined in (34).
Proof
It follows from (41), by virtue of (33) and (40). \(\square \)
Note that, by taking the limit as \(x\downarrow 0\) in (42), we obtain (33).
Remark 2
(Proportional case) We assume that (14) holds. We prove that the second of (16) follows from (42).
Indeed, recalling (36) and (38), from (42) for \(t\ge t_0\), \(x>0\) and \(x_0>0\) one has:
By virtue of (17), one obtains
for \(a>0\), \(\xi <2\) and \(x_0>0\). Hence, under the assumption (14), Eq. (43) leads to the second of (16). \(\Diamond \)
4 The First-Passage Time Through the Zero-State
We now focus on the distribution function of the FPT through the zero-state for the time-inhomogeneous Feller-type diffusion process X(t), with infinitesimal moments (1), when \(\alpha (t)\), \(\beta (t)\) and r(t) are continuous functions such that \(\alpha (t)\in {\mathbb {R}}\), \(\beta (t)\in {\mathbb {R}}\), \(r(t)>0\), \(\beta (t)\le \xi \, r(t)\), with \(0\le \xi <1\). The FPT problem of X(t) through the zero-state can be studied starting from Eq. (8) and making use of (42).
Proposition 8
Under the assumption of Proposition 1, for \(t\ge t_0\) and \(x_0>0\) one has:
with \(R(t|t_0)\) defined in (13), \(B_n(d_1,d_2,\ldots ,d_n)\) given in (31) and (32) and where
denotes the confluent hypergeometric function (Kummer’s function).
Proof
The proof is given in Appendix E. \(\square \)
Remark 3
(Proportional case) We assume that (14) holds. We prove that from (44) one obtains (19).
Indeed, recalling (36) and making use of the relation \(\varPhi (1,2;z)=(e^z-1)/z\), from (44) for \(t\ge t_0\) and \(x_0>0\) one has:
Since (Tricomi [45], p. 31, no. 10)
from (46) one has:
The incomplete gamma function (20) can be expressed in terms of the Kummer’s function (cf. Tricomi [45]p. 160, no. 7):
so that Eq. (19) follows from (48). \(\Diamond \)
Relation (44) plays an important role in the determination of the FPT distribution function and of the FPT density through the zero-state. Indeed, by virtue of (8), for \(t\ge t_0\) and \(x_0>0\) the FPT distribution function is
so that the FPT density through the zero-state can be obtained as
Proposition 9
Under the assumption of Proposition 1, for \(t\ge t_0\) and \(x_0>0\) one has:
with \(A(t|t_0)\) and \(R(t|t_0)\) defined in (13) and \(B_n(d_1,d_2,\ldots ,d_n)\) given in (31) and (32).
Proof
The proof is given in Appendix F. \(\square \)
Remark 4
(Proportional case) We assume that (14) holds. We prove that from (51) one obtains (22).
Indeed, from (36) one has \(B_0=1\) and \(B_n(d_1,d_2,\ldots ,d_n)=(\xi )_n\) for \(n=1,2,\ldots \), so that, under the assumption (14), for \(t\ge t_0\) and \(x_0>0\) from (51) one has:
Making use of (47), it results
5 Special Cases
Under the assumption (14), we analyze the cases in which the growth intensity function \(\alpha (t)\), or the immigration intensity function \(\beta (t)\) or both of them have some kind of periodicity. These cases are of interest in various applied fields, such as in population growth and in queueing systems. Indeed, periodic immigration intensity functions play an important role in the description of the evolution of dynamic for systems influenced by seasonal immigration or other regular environmental cycles. Furthermore, periodic growth intensity functions express the existence of fluctuation in the population dynamics and the presence of rush hours occurring on a daily basis in queueing systems.
5.1 Periodic Immigration Intensity Function
We consider the time-inhomogeneous Feller-type process X(t) such that
with \(\alpha \in {\mathbb {R}}\), \(0\le \xi <1\) and
where \(\nu >0\) is the average of the periodic function r(t) of period Q, c is the amplitude of the oscillations, with \(0 \le c < 1\). From (13), for \(t\ge t_0\) one has \(A(t|t_0)=\alpha \,(t-t_0)\) and
Then, from (55) one obtains:
so that, by virtue of (23), the FPT through the zero-state is a certain event for \(\alpha \le 0\). Moreover, for \(\alpha =0\) the FPT moments (26) are divergent.
In Figs. 1, 2 and 3, the FPT distribution \(G(0,t|x_0,t_0)=1-\int _0^{+\infty }f_a(x,t|x_0,t_0)\,dx\), obtained making use of (19), and the FPT pdf \(g(0,t|x_0,t_0)\), given in (22), are plotted as function of t for the diffusion process (53) for some choices of parameters. In Fig. 4, the mean \(t_1(0|x_0,t_0)\) and the coefficient of variation \(\mathrm{CV}(0|x_0,t_0)\), obtained making use of (26), are plotted as function of \(\nu \) for \(\xi =0, 0.3,0.6\). We note that as \(\nu \) increases, the FPT mean \(t_1(0|x_0,t_0)\) decreases whereas the coefficient of variation increases. Instead, as \(\xi \) increases in [0, 1), the FPT mean increases and the coefficient of variation decreases, due to a raise of the immigration intensity function.
5.2 Periodic Growth Intensity Function
We consider the time-inhomogeneous Feller-type process X(t) such that
with \(r>0\), \(0\le \xi <1\) and
where \(\eta \in {\mathbb {R}}\) is the average of the periodic function \(\alpha (t)\) of period \(Q_1\), b determines the amplitude of the oscillations, with \(0 \le b < 1\). In Fig. 5, the intensity function (57) is plotted as function of t for some choices of parameters \(\eta \), b and \(Q_1\). The dotted lines refer to the average cases, in which \(\alpha (t)=\eta \) with \(\eta =-5\) (bottom) and \(\eta =5\) (top). From (13), for \(t\ge t_0\) one has
and
Then, from (59) one obtains:
so that, by virtue of (23), the FPT through zero-state is a certain event for \(\eta \le 0\). Moreover, for \(\eta =0\) the FPT moments (26) are divergent.
In Fig. 6, the FPT pdf \(g(0,t|x_0,t_0)\), given in (22), is plotted as function of t for the process (56) for some choices of parameters. Instead, in Fig. 7, the mean \(t_1(0|x_0,t_0)\) and the coefficient of variation \(\mathrm{CV}(0|x_0,t_0)\), obtained making use of (26), are plotted as function of r for \(\xi =0, 0.3,0.6\). We note that as r increases, the FPT mean \(t_1(0|x_0,t_0)\) decreases, whereas the coefficient of variation increases. Moreover, the FPT mean and the coefficient of variation increase with \(\xi \) in [0, 1).
5.3 Periodic Immigration and Growth Intensity Functions
We consider the time-inhomogeneous Feller-type process X(t) such that
with \(0\le \xi <1\), r(t) defined in (54) and \(\alpha (t)\) given in (57). Recalling (13), for \(t\ge t_0\) one obtains \(A(t|t_0)\) given in (58) and
The explicit expression of \(R(t|t_0)\) in (61) is obtained in Appendix G. We note that \(\lim _{t\rightarrow +\infty } R(t|t_0)\) diverges as \(\eta \le 0\), so that, due to (23), the FPT through the zero-state is a certain event for X(t).
In Fig. 8, the FPT pdf \(g(0,t|x_0,t_0)\), given in (22), is plotted as function of t for the process (60) for some choices of parameters. Comparing Figs. 6 and 8 , we note the effect of the different periodicities of the growth intensity function \(\alpha (t)\), with \(Q_1=1\), and of the immigration intensity function \(\beta (t)=\xi \,r(t)\), with \(Q=2\). In Fig. 9, the mean \(t_1(0|x_0,t_0)\) and the coefficient of variation \(\mathrm{CV}(0|x_0,t_0)\), obtained making use of (26), are plotted as function of \(\nu \) for \(\xi =0, 0.3,0.6\). As \(\nu \) increases, the FPT mean \(t_1(0|x_0,t_0)\) decreases whereas the coefficient of variation increases. Instead, as \(\xi \) increases in [0, 1), both the FPT mean and the coefficient of variation increase.
6 Concluding Remarks
In this paper, we have considered a time-inhomogeneous Feller-type diffusion process \(\{X(t),t\ge t_0\}\), \(t_0\ge 0\), with infinitesimal drift \(A_1(x,t)=\alpha (t)\,x+\beta (t)\) and infinitesimal variance \(A_2(x,t)=2\,r(t)\,x\), defined in the state-space \([0,+\infty )\), with \(\alpha (t)\in {\mathbb {R}}\), \(\beta (t)\in {\mathbb {R}}\), \(r(t)>0\), \(\beta (t)\le \xi \, r(t)\), with \(0\le \xi <1\), for all \(t\ge t_0\). We have assumed that the zero-state represents an absorbing boundary for X(t). This process plays a relevant role in different fields, including physics, biology, neuroscience, finance and others. For instance, in population biology \(\alpha (t)\) represents the growth intensity function and can be positive, negative or zero at different time instants, \(\beta (t)\) describes the immigration intensity function; instead, the noise intensity function r(t) takes into account the environmental fluctuations. For this process, the transition density \(f_a(x,t|x_0,t_0)\) in the presence of an absorbing boundary in zero-state and the FPT density \(g(0,t|x_0,t_0)\) from \(X(t_0)=x_0\) to the zero-state are obtained. Special attention is dedicated to the proportional case, in which the immigration intensity function and the noise intensity function are related as \(\beta (t)=\xi \,r(t)\), with \(0\le \xi <1\). Various numerical computation are performed to illustrate the effect of periodic intensity functions on the FPT pdf \(g(0,t|x_0,t_0)\), by assuming that \(\alpha (t)\), \(\beta (t)\) or both these functions exhibit some kind of periodicity.
References
Giorno, V., Nobile, A.G.: On the construction on a special class of time-inhomogeneous diffusion processes. J. Stat. Phys. 177(2), 299–323 (2019)
Giorno, V., Nobile, A.G.: Restricted Gompertz-type diffusion processes with periodic regulation functions. Mathematics 7, 555 (2019)
Albano, G., Giorno, V.: Inference on the effect of non homogeneous inputs in Ornstein-Uhlenbeck neuronal modeling. Math. Biosci. Eng. 17(1), 328–348 (2019)
Ghost, H.: Prajneshu: Gompertz growth model in random environment with time-dependent diffusion. J. Stat. Theory Pract. 11, 746–758 (2017)
Buonocore, A., Caputo, L., Nobile, A.G., Pirozzi, E.: Restricted Ornstein-Uhlenbeck process and applications in neuronal models with periodic input signals. J. Comp. Appl. Math. 285, 59–71 (2015)
Gutiérrez, R., Ricciardi, L.M., Román, P., Torres, F.: First-passage-time densities for time-non-homogeneous diffusion processes. J. Appl. Prob. 34(3), 623–631 (1997)
Di Crescenzo, A., Giorno, V., Krishna Kumar, B., Nobile, A.G.: A time-non-homogeneous double-ended queue with failures and repairs and its continuous approximation. Mathematics 6(5), 81 (2018)
Román-Román, P., Serrano-Pérez, J.J., Torres-Ruiz, F.: Fitting real data by means of non-homogeneous lognormal diffusion processes. Stat. Interface 10, 587–600 (2017)
Molini, A., Talkner, P., Katul, G.G., Porporato, A.: First passage time statistics of Brownian motion with purely time dependent drift and diffusion. Physica A 390, 1841–1852 (2011)
Gan, X., Waxman, D.: Singular solution of the Feller diffusion equation via a spectral decomposition. Phys. Rev. E Stat. Nonlinear Soft. Matter Phys. 19(1), 012123 (2015)
Abundo, M.: On the first-passage times of certain Gaussian processes, and related asymptotics. Stoch. Anal. Appl. (2020) https://doi.org/10.1080/07362994.2020.1843495
Feller, W.: Two singular diffusion problems. Ann. Math. 54(1), 173–182 (1951)
Feller, W.: The parabolic differential equations and the associated semi-groups of transformations. Ann. Math. 55, 468–518 (1952)
Peskir, G.: On boundary behaviour of one-dimensional diffusions: from Brown to Feller and beyond. In: Schilling, R.L., Vondrac̆ek, Z., Woyczynski, W.A. (eds.) William Feller, Selected Papers II, Springer, 77–93 (2015)
Karlin, S., Taylor, H.W.: A Second Course in Stochastic Processes. Academic Press, New York (1981)
Giorno, V., Nobile, A.G., Ricciardi, L.M., Sacerdote, L.: Some remarks on the Rayleigh process. J. Appl. Prob. 23(2), 398–408 (1986)
Capocelli, R.M., Ricciardi, L.M.: On the transformation of diffusion processes into the Feller process. Math. Biosci. 29, 219–234 (1976)
Lavigne, F., Roques, L.: Extinction times of an inhomogeneous Feller diffusion process: a PDF approach. Expo. Math. (2020). https://doi.org/10.1016/j.exmath.2019.12.002
Masoliver, J.: Nonstationary Feller process with time-varying coefficients. Phys. Rev. E 93(012122), 1–11 (2016)
Ricciardi, L.M., Di Crescenzo, A., Giorno, V., Nobile, A.G.: An outline of theoretical and algorithmic approaches to first passage time problems with applications to biological modeling. Math. Jpn. 50(2), 247–322 (1999)
Di Crescenzo, A., Nobile, A.G.: Diffusion approximation to a queueing system with time-dependent arrival and service rates. Queueing Syst. 19, 41–62 (1995)
Ditlevsen, S., Lánský, P.: Estimation of the input parameters in the Feller neuronal model. Phys. Rev. E 73(061910), 1–9 (2006)
Lánský, P., Sacerdote, L., Tomassetti, F.: On the comparison of Feller and Ornstein-Uhlenbeck models for neural activity. Biol. Cybern. 73, 457–465 (1995)
Nobile, A.G., Pirozzi, E.: On time non-homogeneous Feller-type diffusion process in neuronal modeling. In: Moreno-Díaz, R. et al. (eds.) EUROCAST 2015, LNCS 9520, 183–191 (2015)
Giorno, V., Lánský, P., Nobile, A.G., Ricciardi, L.M.: Diffusion approximation and first-passage-time problem for a model neuron. III. A birth-and-death process approach. Biol. Cybern. 58(6), 387–404 (1988)
Giorno, V., Nobile, A.G., Ricciardi, L.M.: On neuronal firing modeling via specially confined diffusion processes. Sci. Math. Jpn. 58(2), 265–294 (2003)
Buonocore, A., Giorno, V., Nobile, A.G., Ricciardi, L.M.: A neuronal modeling paradigm in the presence of refractoriness. BioSystems 67, 35–43 (2002)
Tian, Y., Zhang, H.: Skew CIR process, conditional characteristic function, moments and bond pricing. Appl. Math. Comput. 329, 230–238 (2018)
Cox, J.C., Ingersoll, J.E., Jr., Ross, S.A.: A theory of the term structure of interest rates. Econometrica 53, 385–407 (1985)
Linetsky, V.: Computing hitting time densities for CIR and OU diffusions. Applications to mean-reverting models. J. Comput. Financ. 7, 1–22 (2004)
Göing-Jaeshke, A., Yor, M.: A survey and some generalizations of Bessel processes. Bernoulli 9(2), 313–349 (2003)
Giorno, V., Nobile, A.G.: Bell polynomial approach for time-inhomogeneous linear birth-death process with immigration. Mathematics 8, 1123 (2020)
Kolmogoroff, A.: Über die analytischen Methoden in der Wahrscheinlichkeitsrechnung. Math. Ann. 104, 415–458 (1931). https://doi.org/10.1007/BF01457949
Dynkin, E.B.: Kolmogorov and the theory of Markov processes. Ann. Prob. 17(3), 822–832 (1989)
Masoliver, J., Perelló, J.: First-passage and escape problems in the Feller process. Phys. Rev. E 86, 041116 (2012)
Buonocore, A., Caputo, L., Nobile, A.G., Pirozzi, E.: On some time-non-homogeneous linear diffusion processes and related bridges. Sci. Math. Jpn. 76(1), 55–77 (2013)
D’Onofrio, G., Lánský, P., Pirozzi, E.: On two diffusion neuronal models with multiplicative noise: the mean first-passage time properties. Chaos 28, 043103 (2018)
Giorno, V., Nobile, A.G., Pirozzi, E., Ricciardi, L.M.: On the construction of first-passage-time densities for diffusion processes. Sci. Math. Jpn. 64(2), 277–298 (2006)
Giorno, V., Nobile, A.G., Pirozzi, E., Ricciardi, L.M.: FPT densities constructions from Ornstein-Uhlenbeck process. In: Trappl, R. (ed.) Cybernetics and Systems, pp. 244–249. Austrian Society for Cybernetic Studies, Vienna, Austria (2008)
Albano, G., Giorno, V.: On short-term loan interest rate models: a first passage time approach. Mathematics 6, 70 (2018)
Di Nardo, E., D’Onofrio, G.: A cumulant approach for the first-passage-time problem of the Feller square-root process. Appl. Math. Comput. 391, 125707 (2021)
Abramowitz, I.A., Stegun, M.: Handbook of Mathematical Functions. Dover Publications, New York (1972)
Erdèlyi, A., Magnus, W., Oberthettinger, F., Tricomi, F.G.: Tables of Integral Transforms, vol. 1. Mc Graw-Hill, New York (1954)
Erdèlyi, A., Magnus, W., Oberthettinger, F., Tricomi, F.G.: Higher Trascendental Functions, vol. II. Mc Graw-Hill, New York (1953)
Tricomi, F.G.: Funzioni ipergeometriche confluenti. Monografie Matematiche a cura del Consiglio Nazionale delle Ricerche. Edizioni Cremonese, Roma (1954)
Williams, W.E.: Partial Differential Equations. Clarendon Press, Oxford (1980)
Comtet, L.: Advanced Combinatorics: The Art of Finite and Infinite Expansions. D. Reidel Publishing Company, Dordrecht (1974)
Gradshteyn, I.S., Ryzhik, I.M.: Table of Integrals. Series and Products, Academic Press Inc, New York (2014)
Acknowledgements
This research is partially supported by MIUR - PRIN 2017, Project “ Stochastic Models for Complex Systems”. The authors are members of the group GNCS of INdAM.
Funding
: Open access funding provided by University degli Studi di Salerno within the CRUI-CARE Agreement.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Christian Maes.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix
Proof of Proposition 1
To solve (10) with initial condition (11), we use the method of characteristics (cf., for instance, Williams [46]) and we consider the following differential equations:
with the initial conditions:
The first equation of (A1), with the related initial condition in (A2), leads to \(t_0=\xi \). Then, solving the second equation in (A1) with \(t_0=\xi \) and making use of the second of (A2), one has:
Moreover, solving the third equation in (A1) with \(t_0=\xi \) and s given in (3), we have
where the third of (A2) has been used. From (3) with \(\xi =t_0\), we also obtain
Hence, recalling that \(\xi =t_0\) and making use of (5), from (4) it follows:
Since
one has:
We note that
being
Making use of (7) and (8) in (6), one obtains (12). Finally, we note that the assumptions on the functions \(\alpha (t), \beta (t)\) and r(t) in Proposition 1 imply that
so that
i.e. the condition (5) is satisfied. \(\square \)
Proof of Proposition 2
We note that
where the last identity follows by virtue of (9). Hence, making use of (B1) in (12), one obtains (15). To derive (16), we consider the inverse Laplace transform of (15) distinguishing two cases: (i) \(x=0\) and (ii) \(x>0\).
Case (i) If \(x=0\), Eq. (15) becomes:
Since (cf. Erdèlyi et al. [43], p. 144, no. 3)
taking the inverse Laplace transform in (B2), for \(t\ge t_0\) the first of (16) immediately follows.
Case (ii) Let \(x_0>0\). By setting
in (15), making use of (9), one has:
Since (cf. Erdèlyi et al. [43], p. 197, no. 18)
taking the inverse Laplace transform in (B4), for \(t\ge t_0\) one obtains:
from which, applying again the transformation \(x_0=R(t|t_0)\,y\), the second of (16) follows. \(\square \)
Proof of Proposition 5
Let \(x_0>0\) and \(t\ge t_0\). Making use of (B3) in (30) and recalling (9), one has:
We note that
where the last equality follows being
Since (cf., for instance, Comtet [47]):
where \(B_n(d_1,d_2,\ldots ,d_n)\) are the complete Bell polynomials defined in (31), with \(d_k\) given in (32), from (C2) one obtains:
Then, making use of (C3) in (C1) one has:
Finally, since (cf. Gradshteyn and Ryzhik [48], p. 809, no. 8)
by setting \(\alpha =1\), Eq. (C4) leads to:
Applying again the transformation \(x_0=R(t|t_0)\,y\), one obtains (33). \(\square \)
Proof of Proposition 6
We use (B3) in (29), so that, by virtue of (39), for \(t\ge t_0\) we obtain:
Since (cf. Erdèlyi et al. [43], p. 197, no. 16)
from (D1) for \(t\ge t_0\) and \(x>0\) one has:
Then, applying the transformation \(x_0=R(t|t_0)\,y\), Eq. (40) follows from (D2), recalling that \(\delta (a\,x)=\delta (x)/|a|\) and \(g(x)\,\delta (x-a)=g(a)\,\delta (x-a)\). \(\square \)
Proof of Proposition 8
From (42), we obtain:
We note that (cf. Erdèlyi et al [43], p. 197, no. 18)
so that from (E1), by virtue of (34), it follows:
Recalling the expression of the Laguerre polynomials (35), one has:
Then, making use of (E3) in (E2), for \(t\ge t_0\) and \(x_0>0\) one obtains:
Moreover, since the Laguerre polynomials satisfy the following functional relations (cf. Gradshteyn and Ryzhik [48], p. 1001, no. 8.971.4 and no. 8.971.5)
one also has:
Hence, (E4) can be rewritten as
Finally, since (cf. Gradshteyn and Ryzhik [48], p. 1001, no. 8.972.1)
Eq. (44) follows immediately from (E5). \(\square \)
Proof of Proposition 9
Making use of (44) in (50), one has
Since (cf. Gradshteyn and Ryzhik [48], p. 1023, no. 9213)
one obtains:
Therefore, Eq. (F1) can be rewritten as:
where the use of the following relations
has been made. Finally, recalling that
the expression in square bracket in Eq. (F2) becomes:
where the last identity follows from (F3). Then, substituting (F4) in (F2), we obtain Eq. (51). \(\square \)
Evaluation of \(R(t|t_0)\) in (61)
From (61) one has:
with \(0 \le b < 1\) and \(0 \le c < 1\), where
and
for \(Q\ne Q_1\), whereas
for \(Q=Q_1\). \(\square \)
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Giorno, V., Nobile, A.G. Time-Inhomogeneous Feller-type Diffusion Process with Absorbing Boundary Condition. J Stat Phys 183, 45 (2021). https://doi.org/10.1007/s10955-021-02777-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10955-021-02777-3