MDP modeling for multi-stage stochastic programs

ArXi:2509.22981v2 Announce Type: replace We study a class of multi-stage stochastic programs, which incorporate modeling features from Marko decision processes (MDPs). This class includes structured MDPs with continuous action and state spaces. We extend policy graphs to include decision-dependent uncertainty for one-step transition probabilities as well as a limited form of statistical learning. We focus on the expressiveness of our modeling approach, illustrating ideas with a series of examples of increasing complexity.