... Decision Process Single-Agent: basic MDP, SMDP, HSMDP Mobile Ad Hoc Network Multi-Agent: DEC-POMDP How to Achieve Stability? How to Control & Optimize? How to Derive Bounds? Model-Based: DP Irreducibility, ... V-Uniform Ergodicity, Drift Conditions Model-Free: NDP/RL, Policy Gradient Main Contribution: Achieve Stability, Bounds, Optimal Control & Optimization simultaneously ệ ẵẵ è ì ì ẵẵ ểềỉệ Use Stability ... Geometric Drift Conditions Other Contributions: Decentralized Model-free Algorithm, Handle Partial Observability & Locality of Interaction ỉ ểềì ề ỉ ểềá ỉ ểềỉệểéé ểẹ ỉệ èể ỉ ệìỉ ìỉ é ỉí ề ề ề ỉ ệ ỉ...