Two-Sided First Exit Problem for Jump Diffusion Distribution Processes Having Jumps with a Mixture of Erlang ()
1. Introduction
Consider the following jump diffusion process
(1.1)
where the constant
is the starting point of
,
and
represent the drift and the volatility of the diffusion part, respectively,
is a standard Brownian motion with
,
is a Poisson process with rate
, and the jumps sizes
are assumed to be i.i.d. real valued random variables with common density
. Moreover, it is assumed that the random processes
,
and random variables
are mutually independent. In this paper we are interested in the density
of following type
(1.2)
where
,
,
,
and that
,
for all
. Moreover,

Define
to be the first exit time of
to two flat barriers
and
, i.e.

Recently, one-sided and two-sided first exit problems for processes with two-sided jumps have attracted a lot of attentions in applied probability (see [1-7]). For example, Perry and Stadje [1] studied two-sided first exit time for processes with two-sided exponential jumps; Kou and Wang [2] studied the one-sided first passage times for a jump diffusion process with exponential positive and negative jumps. Cai [3] investigated the first passage time of a hyper-exponential jump diffusion process. Cai et al. [4] discussed the first passage time to two barriers of a hyper-exponential jump diffusion process. Closed form expressions are obtained in Kadankova and Veraverbeke [5] for the integral transforms of the joint distribution of the first exit time from an interval and the value of the overshoot through boundaries at the exit time for the Poisson process with an exponential component. For some related works, see Perry et al. [8], Cai and Kou [9], Lewis and Mordecki [10] and the references therein.
Motivated by works mentioned above, the main objective of this paper is to study the first exit time of the process (1.1) with jump density (1.2) from an interval and the overshoot over the boundary at the exit time. In Section 2, we study the roots of the generalized Lundberg equation and conditional memory lessness. The main results of this paper are given in Section 3.
2. Preliminary Results
It is easy to see that the infinitesimal generator of
is given by

for any twice continuously differentiable function
and the Lévy exponent of
is given by

By analytic continuation, the function
can be extended to the complex plane except at finitely many poles. In the following, we consider the resulting extension
of
, i.e.,

Let us denote
and
.
In [11], Kuznetsov has studied the roots of the equation
. However, for this particular Lévy process
, we will give another simple proof for the roots of this equation.
Lemma 2.1. For fix
, the generalized CramérLundberg equation

has
complex roots
with
for
and
with
for
.
Proof. Let


Firstly, we prove that for given
,
has
roots with negative real parts. Set
with
where
is an arbitrary positive constant. Applying Rouchés theorem on the semi-circle
, consisting of the imaginary axis running from
to
and with radius
running clockwise from
to
. We let
and denote by
the limiting semi-circle. It is known that both
and
are analytic in
. We want to show that

Notice that
for
, and
is bounded for
. Hence, for
,

on the boundary of the half circle in
. For
, we have
(see Lewis and Mordecki [10]). On the other hand,

Thus we have
. Since
has
roots with negative real parts, so equation
has
roots with negative real parts. Similarly, we can prove
has
roots with positive real parts.
In the rest of this paper, we assume all the roots of equation
are distinct and denote
,
for notational simplicity, and denote
(or
in the sequel) representing the expectation (or probability) when
starts from
. We denote a sequence of events
, 
= {
:
crosses
at time
by the
th phase of
th positive jump whose parameter is
},
= {
:
crosses
at time
by the
th phase of
th negative jump whose parameter is
}
for
,
,
,
,
and
.
Theorem 2.2. For any
, we have
(2.1)
(2.2)
Furthermore, conditional on
, the stopping time
is independent of the overshoot
(the undershoot
). More precisely, for any
, we have
(2.3)
(2.4)
Proof. Firstly, we prove (2.1) and (2.3). It suffices to show
(2.5)
since (2.1) can be obtained by letting
in (2.5) and then dividing both sides of the resulting equation by
. It is known that an Erlang(n) random variable can be expressed as an independent sum of
exponential random variables with same parameters. Let
the
independent exponentially distributed random variables with parameter
. Denote by
the arrival times of the Poisson process
, and let
be the field generated by process
,
. It follows that

With
, we have

Thus we have

(2.2) and (2.4) can be obtained similarly. This completes the proof.
The following results are immediate consequences of Theorem 2.2.
Corollary 2.3. For
,
,
,
,
,
, we have


Corollary 2.4. For any
, we have




where


for
,
,
,
.
Corollary 2.5. For
,
,
,
, we have


Remark 2.6. When
,
, (2.1) and (2.3) reduce to Equations (8) and (9) of Cai [3], respectively.
3. Main Results
In this section, we study the distribution of the first exit problem to two barriers. We first define three vectors:



where


Let




Define a matrix
.
Theorem 3.1. Consider any nonnegative measurable function
such that
and
for
,
,
,
. For any
and
, we have
(3.1)
where
satisfies
(3.2)
Moreover, when
is a non-singular matrix,
is the unique solution of (3.2), i.e.,
(3.3)
Proof. By the law of total probability, we have

It follows from Corollary 2.4, for
,
,
,
, we have




Combining these equations, we get

The expressions for
,
,
and
can be determined as follows. Let
denote the set of functions 
such that
is twice continuously differentiable and bounded for
with
and
bounded for
. By applying Itô formula to the process
, we have that for
and
,

where
is a martingale with
. Note that we have
as
.
For any
, we can easily obtain from the above equation that

where the last term of the above equation is a mean-0 martingale. This implies that
(3.4)
By simple calculation, the function
with
and
satisfies
for
. It follows from (3.4) that the process
is a martingale. Then
(3.5)
Setting
for
and
for
in (3.5), we have the following linear equations:

and

Then the vector
satisfies
and we have (3.1). If
is non-singular, we have
. This completes the proof.
Corollary 3.2. For any
, we have
(3.6)
where

and

Remark 3.3. When
,
, (3.1) and (3.6) reduce to equation (6) and (15) of [4], respectively.
From Theorem 3.1, choosing
to be
,
,
,
,
,
and
respectively, we can obtain the following corollaries.
Corollary 3.4. 1) For any
, we have
(3.7)
where

is determined by the linear system
. Here

2) For any
, we have
(3.8)
where

is determined by the linear system
. Here

Corollary 3.5. 1) For
and any
,
, we have
(3.9)
where

is determined by the linear system
. Here

2) For
and any
,
, we have
(3.10)
where

is determined by the linear system
. Here

Note that the difference of
and
is exactly
. Thus we obtain the following results.
Corollary 3.6. 1) For
, and for any
, we have
(3.11)
where

is determined by the linear system
. Here

2) For
and any
,
, we have
(3.12)
where

is determined by the linear system
. Here

To end the paper, we give an example.
Example 3.7. When
,
and
, the equation
has
real roots:
,
,
and
. Let

Denote
by

Then we have

where












We define
(
,
,
) and
(
,
,
) as follows: let
(
,
,
) be obtained from
(
,
,
) by changing
to
in
(
,
); let
(
,
,
) be obtained from
(
,
,
) by changing
to
in
(
,
,
).
• If
, then we have

where




• 
• If
, then we have

where

• 
• If
, then we have

where

• 
• If
,
, then we have


• 
• If
,
, then we have

where

• 
• If
, then we have

where

• 
• If
, then we have

where


When
, we have







Therefore, we have



These results are all consistent with that of Theorem 3.1 of Kou and Wang [2] for the one-sided exit problem of the doubly exponential jump diffusion process.
NOTES
#Corresponding author.