Abstract:Dating from the work of Neuts in the 1980s, the field of matrix-analytic methods has been developed to analyse discrete or continuous-time Markov chains with a two-dimensional state space in which the increment of a level variable is governed by an auxiliary phase variable. More recently, matrix-analytic techniques have been applied to general Markov additive models with a finite phase space. The basic assumption underlying these developments is that the process is skip-free (in the case of QBDs or fluid queues) or that it is one-sided, that is it is jump-free in one direction.
From the Markov additive perspective, traditional matrix-analytic models can be viewed as special cases: for M/G/1 and GI/M/1-type Markov chains, increments in the level are constrained to be lattice random variables and for fluid queues, they have to be piecewise linear.
In this paper we discuss one-sided lattice and non-lattice Markov additive processes in parallel. Results that are standard in one tradition are interpreted in the other, and new perspectives emerge. In particular, using three fundamental matrices, we address hitting, two-sided exit, and creeping probabilities.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is related to the analysis of a special class of Markov Additive Processes (MAPs), namely, the one - sided jump processes with lattice and non - lattice increments. Specifically, the author aims to synthesize and extend the existing matrix - analytic methods to deal with problems such as the first hitting time, two - sided exit probabilities, and climbing probabilities in these processes.
### Problem Background
1. **Markov Additive Processes (MAPs)**
- MAPs are a bivariate Markov process \((X_t, J_t)\), where \(X_t\) is called the level component and \(J_t\) is called the phase component.
- The characteristic of these processes is that the increments of the level component \(X_t\) are controlled by the phase component \(J_t\), and the phase component \(J_t\) follows the evolution law of a continuous - time Markov chain.
2. **Limitations of Existing Research**
- Traditional matrix - analytic methods are mainly applied to discrete or continuous - time Markov chains, and usually assume that the process is jump - free (such as QBD or fluid queues) or one - sided jump (i.e., there are no jumps in one direction).
- For MAPs with lattice and non - lattice increments, the results in the existing literature are usually discussed separately, lacking a unified perspective.
### Research Objectives
1. **Unified Framework**
- The author hopes to unify the results of MAPs with lattice and non - lattice increments by introducing three basic matrices (G, H, R), providing a more comprehensive understanding.
2. **New Perspective and Explanation**
- Regard the traditional matrix - analytic model as a special case, re - interpret the standard results, and discover new insights from them.
- In particular, the author focuses on how to use these basic matrices to deal with problems such as the first hitting time, two - sided exit probabilities, and climbing probabilities.
3. **Application and Promotion**
- Explore the application of these theories in practical problems, such as calculating extreme value distributions, the probability of returning to zero, etc.
### Summary of Mathematical Formulas
- **Basic Matrix Equation**
\[
F(\alpha)=Q\circ(E[e^{\alpha U_{ij}}])+\Delta[\psi_1(\alpha),\ldots,\psi_N(\alpha)]
\]
where \(\circ\) represents element - by - element matrix multiplication, and \(\psi_i(\alpha)=\log E[e^{\alpha X(i)_1}]\) is the Laplace exponent of the Lévy process \(X(i)\).
- **First Hitting Time Matrix \(G\)**
\[
G \text{ satisfies } F(G) = 0
\]
- **Occupancy Time Matrix \(H(x)\)**
\[
H(x)_{ij}=E_i[L_j(x,\infty)]
\]
- **Return - to - Zero Matrix \(R\)**
\[
R(x)=e^{Rx}
\]
where \(R\) is the left - hand solution of the equation \(F(\cdot)=0\).
Through these formulas, the author can systematically analyze and solve the above - mentioned problems, providing a solid theoretical basis for further research in related fields.