Arena: A General Evaluation Platform and Building Toolkit for Multi-Agent Intelligence

By Yuhang Song, Jianyi Wang, Thomas Lukasiewicz, Zhenghua Xu, Mai Xu, Zihan Ding, Lianlong Wu. 2019

37 citations recorded by the source corpus.

Learning agents that are not only capable of taking tests, but also innovating is becoming a hot topic in AI. One of the most promising paths towards this vision is multi-agent learning, where agents act as the environment for each other, and improving each agent means proposing new problems for others. However, existing evaluation platforms are either not compatible with multi-agent settings, or limited to a specific game. That is, there is not yet a general evaluation platform for research on multi-agent intelligence. To this end, we introduce Arena, a general evaluation platform for multi-agent intelligence with 35 games of diverse logics and representations. Furthermore, multi-agent intelligence is still at the stage where many problems remain unexplored. Therefore, we provide a building toolkit for researchers to easily invent and build novel multi-agent problems from the provided game set based on a GUI-configurable social tree and five basic multi-agent reward schemes. Finally, we provide Python implementations of five state-of-the-art deep multi-agent reinforcement learning baselines. Along with the baseline implementations, we release a set of 100 best agents/teams that we can train with different training schemes for each game, as the base for evaluating agents with population performance. As such, the research community can perform comparisons under a stable and uniform standard. All the implementations and accompanied tutorials have been open-sourced for the community at https://sites.google.com/view/arena-unity/.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Acme: A Research Framework for Distributed Reinforcement Learning (2020, background): “Arena (Song et al., 2019) is targeted at expressing multi-agent reinforcement learning.”
A survey and critique of multiagent deep reinforcement learning (2018, methodology): “• Arena [303] is platform for multiagent research44 based on the Unity engine [304].”
SAPIEN: A SimulAted Part-Based Interactive ENvironment (2020, background): “Arena [47], a platform that supports multi-agent environments, is built on top of Unity [23].”
Unity: A General Platform for Intelligent Agents (2018, methodology): “Finally, a set of benchmarks called “Arena” have been built using the Unity ML-Agents Toolkit which speciﬁcally focuses on multi-agent scenarios (Song et al., 2020).”
Multi-agent deep reinforcement learning: a survey (2021, methodology): “Evolutionary algorithms have been used in versatile contexts of multi-agent RL, e.g. for building intrinsic motivation (Wang et al. 2019), shaping rewards (Jaderberg et al. 2019), generating curricula (Long et al. 2020) and analyzing dynamics (Bloembergen et al. 2015).”
AllenAct: A Framework for Embodied AI Research (2020, background): “Arena [44] provides a multi-agent platform for games.”
Dealing with Non-Stationarity in MARL via Trust-Region Decomposition (2021, background): “For a cooperative POSG, we quote the deﬁnition in Song et al. (2020), where i and i (cid:48) are a pair of agents in agent space I ; π i and π i (cid:48) are the corresponding policies in the policy space Π i and Π i (cid:48) respectively.”
A review of cooperative multi-agent deep reinforcement learning (2019, background): “Arena [122] extends the Unity engine and provides an speciﬁc platform to deﬁne and build new multi-agent games and scenarios based on the available games.”
Agentic AI: Autonomous Intelligence for Complex Goals—A Comprehensive Survey (2025, unknown): “Training or evaluating agents in the simulated environment is done using OpenAI Gym [28] and Unity ML-Agents [29].”