A new class of stochastic processes with great potential for interesting applications
Fulgence Eyi Obiang,Paule Joyce Mbenangoya,Magloire Yorick Nguema MBA,Octave Moutsinga
DOI: https://doi.org/10.48550/arXiv.2210.01524
2022-10-04
Abstract:This paper contributes to the study of a new and remarkable family of stochastic processes that we will term class $\Sigma^{r}(H)$. This class is potentially interesting because it unifies the study of two known classes: the class $(\Sigma)$ and the class $\mathcal{M}(H)$. In other words, we consider the stochastic processes $X$ which decompose as $X=m+v+A$, where $m$ is a local martingale, $v$ and $A$ are finite variation processes such that $dA$ is carried by $\{t\geq0:X_{t}=0\}$ and the support of $dv$ is $H$, the set of zeros of some continuous martingale $D$. First, we introduce a general framework. Thus, we provide some examples of elements of the new class and present some properties. Second, we provide a series of characterization results. Afterwards, we derive some representation results which permit to recover a process of the class $\Sigma^{r}(H)$ from its final value and of the honest times $g=\sup\{t\geq0:X_{t}=0\}$ and $\gamma=\sup{H}$. In final, we investigate an interesting application with processes presently studied. More precisely, we construct solutions for skew Brownian motion equations using stochastic processes of the class $\Sigma^{r}(H)$.
Probability