MPE (Multi-Agent Particle Environments)

E438350

multi-agent reinforcement learning benchmark suite research tool software environment collection

MPE (Multi-Agent Particle Environments) is a classic collection of lightweight 2D multi-agent reinforcement learning benchmark environments featuring simple particle-based agents and tasks like cooperation, competition, and communication.

Try in SPARQL Jump to: Surface forms Disambiguation Statements Elicitation Referenced by

All labels observed (3)

Label	Occurrences
MPE (Multi-Agent Particle Environments) canonical	1
Multi-Agent Particle Environments	1
paper "Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments"	1

How this entity was disambiguated

This entity first appeared as the object of triple T4425216 — resolving that mention is where its identity was fixed. The disambiguator weighed these candidate entities and picked the highlighted one (or “None”, minting a new entity). This is how homonymy is resolved: the same surface form can point to different entities.

NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07

Target entity: MPE (Multi-Agent Particle Environments)
Context triple: [PettingZoo, hasEnvironmentType, MPE (Multi-Agent Particle Environments)]

A. Arcade Learning Environment
Arcade Learning Environment is a widely used research platform that provides a suite of Atari 2600 games for developing and evaluating reinforcement learning algorithms.
B. OpenAI Baselines
OpenAI Baselines is a collection of high-quality reference implementations of reinforcement learning algorithms released by OpenAI for research and benchmarking.
C. MuJoCo environments
MuJoCo environments are physics-based continuous control simulation tasks widely used in reinforcement learning research and benchmarking.
D. Dueling DQN
Dueling DQN is a deep reinforcement learning algorithm that separates state-value and advantage estimations within its neural network architecture to improve learning efficiency and stability over standard DQN.
E. DDPG
DDPG (Deep Deterministic Policy Gradient) is a model-free, off-policy deep reinforcement learning algorithm designed for continuous action spaces, combining ideas from DQN and actor-critic methods.
F. None of above. chosen
G. Unsure - the case is ambiguous/there is not enough information to decide.

NED2 Entity disambiguation (via description) gpt-5-mini-2025-08-07

Target entity: MPE (Multi-Agent Particle Environments)
Target entity description: MPE (Multi-Agent Particle Environments) is a classic collection of lightweight 2D multi-agent reinforcement learning benchmark environments featuring simple particle-based agents and tasks like cooperation, competition, and communication.

A. Arcade Learning Environment
Arcade Learning Environment is a widely used research platform that provides a suite of Atari 2600 games for developing and evaluating reinforcement learning algorithms.
B. OpenAI Baselines
OpenAI Baselines is a collection of high-quality reference implementations of reinforcement learning algorithms released by OpenAI for research and benchmarking.
C. MuJoCo environments
MuJoCo environments are physics-based continuous control simulation tasks widely used in reinforcement learning research and benchmarking.
D. Dueling DQN
Dueling DQN is a deep reinforcement learning algorithm that separates state-value and advantage estimations within its neural network architecture to improve learning efficiency and stability over standard DQN.
E. DDPG
DDPG (Deep Deterministic Policy Gradient) is a model-free, off-policy deep reinforcement learning algorithm designed for continuous action spaces, combining ideas from DQN and actor-critic methods.
F. None of above. chosen

Statements (50)

Predicate	Object
instanceOf	multi-agent reinforcement learning benchmark suite ⓘ research tool ⓘ software environment collection ⓘ
acronym	MPE ⓘ
agentActionSpace	continuous control actions ⓘ
agentType	simple point-mass particles ⓘ
associatedAlgorithm	MADDPG ⓘ
complexityLevel	low-dimensional ⓘ
designGoal	simplicity for rapid experimentation ⓘ standardized multi-agent RL benchmarks ⓘ
domain	multi-agent reinforcement learning ⓘ reinforcement learning ⓘ
environmentType	discrete-time simulation ⓘ
feature	continuous 2D space ⓘ lightweight 2D environments ⓘ particle-based agents ⓘ
fullName	Multi-Agent Particle Environments ⓘ linked to: MPE (Multi-Agent Particle Environments)
hasStateElements	agent positions ⓘ agent velocities ⓘ landmark positions ⓘ
implementationLanguage	Python ⓘ
includesEnvironment	simple_adversary ⓘ simple_crypto ⓘ simple_push ⓘ simple_reference ⓘ simple_speaker_listener ⓘ simple_spread ⓘ simple_spread_comm ⓘ simple_tag ⓘ simple_world_comm ⓘ
license	MIT License ⓘ
observationType	partial observations ⓘ
originallyReleasedWith	paper "Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments" ⓘ linked to: MPE (Multi-Agent Particle Environments)
rewardStructure	competitive rewards ⓘ cooperative rewards ⓘ mixed rewards ⓘ
supportsAgentHeterogeneity	true ⓘ
supportsCommunicationChannels	true ⓘ
supportsFramework	OpenAI Gym-style interface ⓘ
supportsTaskType	communication ⓘ competition ⓘ cooperation ⓘ mixed cooperative-competitive tasks ⓘ
typicalNumberOfAgents	multiple agents per environment ⓘ
typicalUseContext	academic research ⓘ
usedFor	benchmarking multi-agent RL algorithms ⓘ evaluating centralized training with decentralized execution ⓘ evaluating decentralized policies ⓘ studying coordination between agents ⓘ studying emergent communication ⓘ

How these facts were elicited

Referenced by (3)

Full triples — surface form annotated when it differs from this entity's canonical label.

PettingZoo → hasEnvironmentType → MPE (Multi-Agent Particle Environments) ⓘ

MPE (Multi-Agent Particle Environments) → fullName → Multi-Agent Particle Environments ⓘ

linked to: MPE (Multi-Agent Particle Environments)

MPE (Multi-Agent Particle Environments) → originallyReleasedWith → paper "Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments" ⓘ