Kavin R,Ananta K. Majee
Abstract:In this article, we study a nonlinear stochastic control problem perturbed by multiplicative Levy noise, where the nonlinear operator in divergence form satisfies p type growth with coercivity assumptions. By using Aldous tightness criteria and Jakubowski version of the Skorokhod theorem on nonmetric spaces along with standard contraction method, we establish existence of pathwise unique strong solution. Formulating the associated control problem, and using variational approach together with the convexity property of cost functional in control variable, we establish existence of a weak optimal solution of the underlying problem. We use the technique of Maslowski and Seidler to prove existence of an invariant measure for uncontrolled SPDE driven with multiplicative Levy noise.
What problem does this paper attempt to address?
The problems that this paper attempts to solve mainly focus on three aspects:
1. **Existence of strong solutions**: The paper aims to prove that for the nonlinear stochastic partial differential equation (SPDE) perturbed by Lévy noise, there exists a path - unique strong solution. Here, the nonlinear operator satisfies a growth condition and coercivity assumption of a certain form.
2. **Existence of weakly optimal controls**: Through variational methods and the convexity property of the cost functional, the paper also explores the existence of weakly optimal solutions in the control problem related to the above - mentioned SPDE. Specifically, given a deterministic objective function and a terminal payoff, the paper looks for a weakly feasible solution such that the cost functional reaches its minimum.
3. **Existence of invariant measures**: For the uncontrolled SPDE, the paper also studies the possibility of the existence of an invariant measure for its strong solution. This involves proving certain properties of the Markov semigroup, such as probability boundedness and the sequential weak Feller property.
### Formula summary
- **SPDE model**:
\[
du-\text{div}_x(A(x, u, \nabla u)+\vec{F}(u))dt = Udt+h(u)dW(t)+\int_E\eta(x, u; z)\tilde{N}(dz, dt)\quad\text{in }\Omega\times D_T,
\]
\[
u = 0\quad\text{on }\Omega\times\partial D_T,
\]
\[
u(0,\cdot)=u_0(\cdot)\quad\text{in }\Omega\times D,
\]
where \(D\subset\mathbb{R}^d\) is a bounded domain with a Lipschitz boundary, \(W(t)\) is a cylindrical Wiener process, and \(\tilde{N}(dz, dt)\) is the compensated Poisson random measure.
- **Conditions for the nonlinear operator**:
- **Coercivity**:
\[
A(x,\lambda,\zeta)\cdot\zeta\geq C_1|\zeta|^p - K_1(x)\quad\text{for a.e. }x\in D\text{ and for all }\zeta\in\mathbb{R}^d,\lambda\in\mathbb{R}.
\]
- **Boundedness**:
\[
|A(x,\lambda,\xi)|\leq C_2|\xi|^{p - 1}+C_3|\lambda|^{p - 1}+K_2(x)\quad\text{for a.e. }x\in D\text{ and for all }\xi\in\mathbb{R}^d,\lambda\in\mathbb{R}.
\]
- **Cost functional**:
\[
J(\pi)=\mathbb{E}\left[\int_0^T\left(\|u(t)-u_{\text{det}}(t)\|_{L^2}^2+\|U(t)\|_{L^2}^2\right)dt+\Psi(u(T))\right].
\]
- **Conditions for the invariant measure**:
\[
C_{hg}+2C_1\|\nabla u\|_{L^p}^p-\|h(u)\|_{L^2(L^2)}^2-\int_E\|\eta(\cdot, u; z)\|_{L^2}^2m(dz)\geq\delta\|u\|_{L^2}^p,
\]
where \(C_{hg}=2(C_