masafumiendo / IntelligentAgentandDecisionMaking

Intelligent Agent and Decision Making

This project aims to implement sequential decision making algorithms w/o and w/ distributed computation. The basic outline of this project is listed as follows:

Basic Markov Decision Process

• Value Iteration (VI) for infinite horizon problems w/o distributed computing
• Value Iteration (VI) for infinite horizon problems w/ distributed computing
• Policy Iteration (PI) for infinite horizon problems

Reinforcement Learning w/ Q Table

• SARSA
• Q-Learning w/o distributed computing
• Q-Learning w/ distributed computing

Reinforcement Learning w/ Function Approximation

• Deep Q-Network w/o distributed computing
• Deep Q-Network w/ distributed computing