AI Paper Atlas

Research map / Multi-agent systems and games / Cooperative multi-agent RL

Deep Reinforcement Learning for Swarm Systems

Maximilian Hüttenrauch, Adrian Šošić, Gerhard Neumann

JMLR 2019 · Multi-agent systems and games › Cooperative multi-agent RL · approach: MARL benchmarks and grouping · 183 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use neural network mean embedding instead of max pooling (ablation)pursuit-evasion policy · Pursuit-evasionsuccessful-catch trials: 4/16 -> 10/16Improved
use mean embedding instead of softmax pooling (ablation)pursuit-evasion policy · Pursuit-evasionsuccessful-catch trials: 6/16 -> 10/16Improved
use neural network embedding NN+ with extended features instead of NN or RBF on basic set (ablation)rendezvous policy · Rendezvousaverage episode return: ~10 % betterImproved
use NN+ embedding instead of NN/RBF (ablation)rendezvous policy · Rendezvoussteady-state mean distance: ~8e-2 -> ~4e-2Improved
use NN+ embedding instead of NN/RBF (ablation)rendezvous policy · Rendezvoustime to reach reference distance: ~25 %Improved
histogram/RBF embedding with extended feature set (ablation)rendezvous policy · Rendezvoussuccess: did not succeedHurt
replace mean embedding with statistics (mean, std, skew, kurtosis) of features (ablation)pursuit-evasion policy · Pursuit-evasion 50 agents 5 evadersreturnHurt
compare learned policy to PD controllerrendezvous policy · Rendezvousaverage rewardImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy