Methods › Reinforcement Learning › Off-Policy TD Control › Double Q-learning

Double Q-learning

112 papers tagged archive 2025-07-28

Introduced by Hado V. Hasselt in Double Q-learning

archive 2025-07-28 Description, source and code snippet are the archive's method entry.

Double Q-learning is an off-policy reinforcement learning algorithm that utilises double estimation to counteract overestimation problems with traditional Q-learning.

The max operator in standard Q-learning and DQN uses the same values both to select and to evaluate an action. This makes it more likely to select overestimated values, resulting in overoptimistic value estimates. To prevent this, we can decouple the selection from the evaluation, which is the idea behind Double Q-learning:

Y^Qₜ = Rₜ₊₁ + γQ(Sₜ₊₁, maxₐQ(Sₜ₊₁, a; θₜ);θₜ)

The Double Q-learning error can then be written as:

Y^(DoubleQ)ₜ = Rₜ₊₁ + γQ(Sₜ₊₁, maxₐQ(Sₜ₊₁, a; θₜ);θ^′ₜ)

Here the selection of the action in the max is still due to the online weights θₜ. But we use a second set of weights θ^′ₜ to fairly evaluate the value of this policy.

Source: Deep Reinforcement Learning with Double Q-learning

PaperSource

Papers archive 2025-07-28

30 shown of 112, newest first. Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code; it is per sample and not a correctness claim.

Tasks archive 2025-07-28

20 shown of 67 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.

TaskPapers
Reinforcement Learning (RL)72
Q-Learning63
reinforcement-learning61
Reinforcement Learning58
Deep Reinforcement Learning47
Atari Games18
OpenAI Gym10
Decision Making9
Continuous Control7
continuous-control6
MuJoCo5
Management4
Multi-agent Reinforcement Learning4
Scheduling4
Efficient Exploration3
General Reinforcement Learning3
Traffic Signal Control3
Computational Efficiency2
Diversity2
Ensemble Learning2

Usage over time archive 2025-07-28

Papers per year tagged with Double Q-learning: 2010 to 2025, peak 27 27 0 2010: 1 paper 2010 2011: 0 papers 2012: 0 papers 2012 2013: 0 papers 2014: 0 papers 2014 2015: 2 papers 2016: 2 papers 2016 2017: 2 papers 2018: 8 papers 2018 2019: 9 papers 2020: 27 papers 2020 2021: 24 papers 2022: 14 papers 2022 2023: 17 papers 2024: 4 papers 2024 2025: 2 papers 2025
Papers per year the archive tags with this method, by the paper's archive date (112 dated). Bars are counts, not a trend claim.

Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).

Categories archive 2025-07-28

Off-Policy TD Control

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections