Asymptotic expansion of the optimal control under logarithmic penalty: worked example and open problems
Abstract
We discuss the problem of expansion of optimal control, state and costate when a logarithmic penalty is applied to constraints. We show that, in a simple case, that the variation of (a regular) junction point, and of the optimal control, state and costate is of order $\eps\log \eps$, where $\eps$ is the penalty parameter.
Origin : Files produced by the author(s)