Compactly Restrictable Metric Policy Optimization Problems

We study policy optimization problems for deterministic Markov decision processes (MDPs) with metric state and action spaces, which we refer to as (MPOPs). Our goal is establish theoretical results on the well-posedness of MPOPs that can characterize practically relevant continuous control systems. To do so, define a special class called compactly restrictable (CR-MPOPs), are flexible enough capture complex behavior robotic systems but specific admit solutions using dynamic programming methods such value iteration. show how arrive at CR-MPOPs forward-invariance. further our be used feedback linearizable affine