Some topics: dynamic programming, Hamilton-Jacobi-Bellman equations, viscosity solutions, dual problems, computational complexity, numerical methods for Hamilton-Jacobi equations, symplectic methods.
Some applications: optimal portfolio (SDE), American options (SDE), catalytic converter (PDE), shape optimization (PDE).
Goal: to understand and be able to use basic mathematical and numerical methods to solve optimal control problems based on differential equations, which includes that the student after the course can:
-
derive the Hamilton-Jacobi-Bellman equation,
-
derive the Pontryagins principle,
-
formulate and derive existence and uniqueness of viscosity solutions,
-
analyze computational complexity for dynamics programming and Lagrange’s method,
-
formulate and analyze numerical methods for Hamilton-Jacobi equations,
-
formulate applications of optimal control, e.g. for inverse problems,
-
analyze and use symplectic numerical methods for Hamiltonian systems.