Variants of Dynamic Programming | Chan`s Jupyter

A summary of “Understanding deep reinforcement learning”

Policy Improvement by Iterative Methods Asynchronous Dynamic Programming In-Place Dynamic Programming Prioritized Sweeping Sample Backups Approximate DP Policy Improvement by Iterative Methods Asynchronous Dynamic Programming Dynamic Programming mentioned before used synchronous backups which updates all staes at each iteration in parallel. (that means, next state value function can be calculated when the current state value function is ready.) Asynchronous DP updates each state in any order. This can significantly reduce computation, and it is convergent if all states continue to…

saved by

• Fernando Silva

related reading

• Policy Gradient Algorithms | Lil'Loglilianweng.github.io
• A (Long) Peek into Reinforcement Learning | Lil'Loglilianweng.github.io
• 4.4 Value Iterationincompleteideas.net
• Is Frontier Asynchronous RL Solved? — Luke J. Huangluk-huang.github.io
• [1707.06347] Proximal Policy Optimization Algorithmsarxiv.org
• Part 3: Intro to Policy Optimization - Spinning Up documentationspinningup.openai.com
• Successor Features for Transfer in Reinforcement Learningproceedings.neurips.cc
• RL_Notes__final_.pdfjubayer-ibn-hamid.github.io
• An Updated Introduction to Reinforcement Learning | Sri's Blogsrianumakonda.com
• Policy gradient methoden.wikipedia.org
• Vanilla Policy Gradient - Spinning Up documentationspinningup.openai.com
• [2602.17616] Stable Asynchrony: Variance-Controlled Off-Policy RL for LLMsarxiv.org