Optimal control for port-Hamiltonian systems and a new perspective on dynamic network flow problems
We formulate open-loop optimal control problems for general port-Hamiltonian systems with possibly state-dependent system matrices and prove their well-posedness. The optimal controls are characterized by the first-order optimality system, which is the starting point for the derivation of an adjoint-based gradient descent algorithm. Moreover, we discuss the relationship of port-Hamiltonian dynamics and minimum cost network flow problems. Our analysis is underpinned by a proof of concept, where we apply the proposed algorithm to static minimum cost flow problems and dynamic minimum cost flow problems with a simple directed acyclic graph. The numerical results validate the approach.
READ FULL TEXT