Abstract:In this paper, we are concerned with a multidimensional backward stochastic differential equation (BSDE) with a general random terminal time $\tau$, which may take values in $[0,+\infty]$. Firstly, we establish an existence and uniqueness result for a weighted $L^p~(p>1)$ solution of the preceding BSDE with generator $g$ satisfying a stochastic monotonicity condition with general growth in the first unknown variable $y$ and a stochastic Lipschitz continuity condition in the second unknown variable $z$. Then, we derive an existence and uniqueness result for a weighted $L^1$ solution of the preceding BSDE under an additional stochastic sub-linear growth condition in $z$. These results generalize the corresponding ones obtained in \cite{Li2024} to the $L^p~(p\geq 1)$ solution case. Finally, the corresponding comparison theorems for the weighted $L^p~(p\geq1)$ solutions are also put forward and verified in the one-dimensional setting. In particular, we develop new ideas and systematical techniques in order to establish the above results.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the existence and uniqueness of weighted \(L^p\) solutions to backward stochastic differential equations (BSDEs) with random monotone generators in a random time - horizon. Specifically, the paper focuses on multi - dimensional BSDEs under a general random stopping time \(\tau\):
\[ Y_t=\xi+\int_{t}^{\tau} g(s, Y_s, Z_s) ds-\int_{t}^{\tau} Z_s dB_s, \quad t\in[0, \tau], \]
where:
- \(\xi\) is an \(F_{\tau}\)-measurable \(k\)-dimensional random vector, called the terminal value.
- \(g(\omega, t, y, z)\) is a random function defined on \(\Omega\times[0, \tau]\times\mathbb{R}^k\times\mathbb{R}^{k\times d}\), called the generator.
The main contributions of the paper are as follows:
1. **Existence and uniqueness of weighted \(L^p (p > 1)\) solutions**: First, the author proves the existence and uniqueness results of weighted \(L^p (p > 1)\) solutions under the conditions that the generator \(g\) satisfies the random monotonicity condition, the general growth condition for the first unknown variable \(y\), and the random Lipschitz continuity condition for the second unknown variable \(z\). This generalizes the corresponding conclusions in Li et al. [2024] to the case of \(L^p\) solutions.
2. **Existence and uniqueness of weighted \(L^1\) solutions**: Secondly, under an additional random sub - linear growth condition, the author also studies the existence and uniqueness of weighted \(L^1\) solutions.
3. **Comparison theorem**: Finally, the paper proposes a comparison theorem for weighted \(L^p (p\geq1)\) solutions and verifies it in the one - dimensional case.
These results not only generalize the existing related work, but also introduce some new methods and techniques to deal with more general integral conditions and weighted spaces. In addition, the paper compares in detail the similarities and differences between the new results and the existing work through specific examples and remarks, showing its innovation.
### Formula presentation
- Random monotonicity condition:
\[ \langle y_1 - y_2, g(\omega, t, y_1, z)-g(\omega, t, y_2, z)\rangle\leq\mu_t(\omega)|y_1 - y_2|^2, \quad t\in[0, \tau], \]
where \(\mu_t(\omega)\) is a non - negative random process.
- Random Lipschitz continuity condition:
\[ |g(\omega, t, y, z_1)-g(\omega, t, y, z_2)|\leq\nu_t(\omega)|z_1 - z_2|, \quad t\in[0, \tau], \]
where \(\nu_t(\omega)\) is a non - negative random process.
- Weighting factor:
For \(p > 1\), use the weighting factor \(e^{\int_{0}^{t}(\beta\mu_s+\rho^2[(p - 1)\wedge1]\nu_s^2)ds}\), where \(\beta\geq1\) and \(\rho > 1\);
For \(p = 1\), use the weighting factor \(e^{\int_{0}^{t}\beta\mu_sds}\) and require \(\int_{0}^{\tau}\nu_t^2dt\leq M\).
These formulas and conditions ensure the existence and uniqueness of solutions to BSDEs in a broader context.