Finite Element Approximation of Hamilton-Jacobi-Bellman equations with nonlinear mixed boundary conditions

Bartosz Jaroszkowski,Max Jensen

DOI: https://doi.org/10.48550/arXiv.2105.09585

2021-05-20

Abstract:We show strong uniform convergence of monotone P1 finite element methods to the viscosity solution of isotropic parabolic Hamilton-Jacobi-Bellman equations with mixed boundary conditions on unstructured meshes and for possibly degenerate diffusions. Boundary operators can generally be discontinuous across face-boundaries and type changes. Robin-type boundary conditions are discretised via a lower Dini derivative. In time the Bellman equation is approximated through IMEX schemes. Existence and uniqueness of numerical solutions follows through Howard's algorithm. Keywords: Finite element method, Hamilton-Jacobi-Bellman equation, Mixed boundary conditions, Fully nonlinear equation, Viscosity solution

Numerical Analysis,Optimization and Control

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: How to use the monotone P1 finite element method to numerically approximate the viscosity solution of the isotropic parabolic Hamilton - Jacobi - Bellman (HJB) equation under the condition of nonlinear mixed boundary conditions, and ensure its uniform convergence under unstructured grids and possibly degenerate diffusion conditions.

Specifically, this research focuses on the following problems:

1. **Equation forms**:
- In the time interval \([0, T)\) and the region \(\Omega\), the form of the HJB equation is:
\[
-\partial_t v+\sup_{\alpha \in A}(L_\alpha v - f_\alpha)=0\quad\text{in }[0, T)\times\Omega
\]
- In the time interval \([0, T)\) and on the boundary \(\partial\Omega_t\), the HJB equation with Robin - type boundary conditions is:
\[
-\partial_t v+\sup_{\alpha \in A}(L_\alpha^{\partial\Omega} v - g_\alpha)=0\quad\text{on }[0, T)\times\partial\Omega_t
\]
- In the time interval \([0, T)\) and on the boundary \(\partial\Omega_R\), the pure Robin - type boundary condition is:
\[
\sup_{\alpha \in A}(L_\alpha^{\partial\Omega} v - g_\alpha)=0\quad\text{on }[0, T)\times\partial\Omega_R
\]
- In the time interval \([0, T)\) and on the boundary \(\partial\Omega_D\), the Dirichlet boundary condition is:
\[
v - g = 0\quad\text{on }[0, T)\times\partial\Omega_D
\]
- At the final time \(T\) and in the region \(\Omega\), the initial condition is:
\[
v - v_T = 0\quad\text{on }\{T\}\times\Omega
\]

2. **Complexity of boundary conditions**:
- The boundary operator \(L_\alpha^{\partial\Omega}\) depends on the entire gradient \(\nabla v\) rather than just the tangential gradient \(\nabla_{\partial\Omega} v\), which means that \(L_\alpha^{\partial\Omega} v\) cannot be calculated only by \(v|_{\partial\Omega}\).
- The Robin - type boundary conditions (1b) and (1c) reflect the possibility of controlling particles or agents on the boundary, while the Dirichlet condition (1d) is applicable to those boundary parts that cannot be controlled but have rewards or costs.

3. **Challenges of numerical methods**:
- It is necessary to discretize the directional derivatives in (1b) and (1c) while maintaining both consistency and monotonicity.
- Due to the change of boundary conditions, the boundary operator may be discontinuous on the face boundary.

4. **Application background**:
- The author is particularly interested in the financial interest rate problem in the Heston model, which raises the problem of mixed boundary conditions under the uncertain market volatility risk price.

In summary, this paper aims to develop a finite element method that can handle the above - mentioned complexity.