Peter Vamplew

A Demonstration of Issues with Value-Based Multiobjective Reinforcement Learning Under Stochastic State Transitions

2020 · arXiv (Cornell University)

We report a previously unidentified issue with model-free, value-based approaches to multiobjective reinforcement learning in the context of environments with stochastic state transitions. An example multiobjective Markov Decision Process (MOMDP) is used to demonstrate that …