Abstract:In this paper, we consider a subclass of piecewise deterministic Markov processes with a Polish state space that involve a deterministic motion punctuated by random jumps, occurring in a Poisson-like fashion with some state-dependent rate, between which the trajectory is driven by one of the given semiflows. We prove that there is a one-to-one correspondence between stationary distributions of such processes and those of the Markov chains given by their post-jump locations. Using this result, we further establish a criterion guaranteeing the existence and uniqueness of the stationary distribution in a particular case, where the post-jump locations result from the action of a random iterated function system with an arbitrary set of transformations.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the existence and uniqueness of the stationary distribution of a class of Piecewise Deterministic Markov Processes (PDMPs) with state - dependent jump intensities. Specifically, the author studied a class of PDMPs defined in Polish spaces. These processes consist of deterministic motion and random jumps, where the jumps occur in the manner of a Poisson process, but their rates depend on the current state. The main objective of the paper is to prove that there is a one - to - one correspondence between the stationary distribution of such processes and the stationary distribution of the Markov chain that describes their post - jump positions, and on this basis, establish a criterion to ensure the existence and uniqueness of the stationary distribution.
### Background of the paper
- **PDMPs**: Piecewise Deterministic Markov Processes are a class of non - diffusive stochastic processes that combine deterministic motion and random jumps. These processes are widely used in many application fields, such as biology, physics, and engineering.
- **State - dependent jump intensity**: Traditional PDMPs assume that the jump intensity is constant, but in this paper, the author considered the case where the jump intensity depends on the current state, which makes the model more flexible and realistic.
### Main contributions
1. **One - to - one correspondence**: The author proved that there is a one - to - one correspondence between the stationary distribution of PDMPs and the stationary distribution of the Markov chain that describes their post - jump positions. Similar results have been obtained in the literature, but usually with the assumption that the jump intensity is constant. This paper extends these results by introducing state - dependent jump intensities.
2. **Existence and uniqueness criteria**: Based on the above one - to - one correspondence, the author further established criteria to ensure the existence and uniqueness of the stationary distribution. This criterion is applicable in the specific case where the post - jump positions are generated by a random iterative function system.
### Mathematical tools
- **Weak generator**: The author used the concept of weak generator, which is a tool for contraction semigroups defined on subspaces of Banach spaces. Through the weak generator, the author was able to handle the deterministic dynamics on non - Euclidean state spaces, which is difficult to handle by traditional methods.
- **Feller property**: The paper assumed a strengthened version of the Feller property, which ensures the joint continuity of certain specific functions, so that the explicit form of the semigroup can be derived.
### Applications
- **Gene expression analysis**: The model in the paper is particularly useful in analyzing the stochastic dynamics of gene expression, especially in the presence of transcriptional bursts. Specifically, the concentration of protein molecules changes over time but is interrupted by randomly occurring production events. From a biological perspective, the frequency of these production events depends on the current number of molecules, so considering non - constant jump intensities makes the model more accurate.
### Conclusion
By proving the one - to - one correspondence between the stationary distribution of PDMPs and the stationary distribution of the Markov chain that describes their post - jump positions, and establishing the existence and uniqueness criteria, this paper provides an important theoretical basis for understanding and analyzing PDMPs with state - dependent jump intensities. These results are not only of great theoretical significance but also have wide applicability in practical applications.