Markovian Decision Processes with Probabilistic Observation of States

This is a study of finite state discrete time discounted Markovian decision process when the states are probabilistically observed. A model of this process is formulated, and an implicit enumeration algorithm is presented which optimizes the total expected discounted reward given the initial state. Several numerical examples are presented.

• 저널Management Science · 20(1) · 1–13
• 토픽Bayesian Modeling and Causal Inference · Artificial Intelligence
• DOI10.1287/mnsc.20.1.1
• 저자Jay Satia, Roy E. Lave