Expected utility maximization beyond the Markovian setting

An overview of the recent approaches used to solve
portfolio optimization problems for general market models
is given.
In particular, the focus will be on dynamic programming
techniques and on their applicability to expected utility
maximization in non-Markovian settings for classical
utilities (power, exponential or log type), including the
case of partial information. Moreover, another method
which works for general utilities is presented and
compared to recent results obtained by dynamic
programming.

This talk is based on joint works with M. Mania, R.
Tevzadze and B. Trivellato.