[agents] Announcing a NEW competition at the AAMAS'2025 Conference: MOASEI

[Leen-Kiat Soh]

Dear Community,

We are excited to announce a new competition, Methods for Open Agent Systems Evaluation Initiative (MOASEI), to be held at the AAMAS'2025 conference in Detroit, MI, USA, in May 2025.  More specifically, the competition will be held leading up to and culminating at the AAMAS conference, providing a unique opportunity for participants to showcase their work to the broader multiagent systems community. We look forward to your participation and hope to see you at the competition

Many real-world applications of multiagent systems (MAS) are open agent systems (OASYS) where the sets of agents and tasks can dynamically change over time. Often, these changes are unpredictable and unknown in advance by the decision-making agents operating to accomplish tasks. In contrast, most methods for autonomous decision making (whether planning, reinforcement learning, or game theory) assume that the set of agents and tasks are static throughout the lifetime of the system. Mismatches between the assumptions of the agents’ reasoning and models of the environment vs. the underlying dynamics of the environment can risk critical failure of agents deployed to real-world applications. In this challenge, competitors will design, train, and submit multiagent reinforcement learning (MARL) solutions to guide agent actions in OASYS domains featuring agent openness (where the set of operating agents changes over time) and task openness (where the set of tasks available to agents change over time).

We will have three separate tracks, each featuring a single simulated domain:

  *
Cybersecurity Defense (Agent Openness only): Two teams of multiple agents (attackers vs. defenders) compete to either infiltrate or protect a network infrastructure. Attacker agents frequently disappear to avoid detection, and defender agents can be taken offline as the equipment they use is disrupted by network infection.
  *
Rideshare (Task Openness only): Agents operating autonomous cars within a ridesharing application decide how to prioritize dynamically appearing passengers as tasks.
  *
Wildfire Suppression (Both Agent and Task Openness): Agents decide how to use limited suppressant resources to collaboratively put out wildfire tasks that appear both spontaneously and due to realistic fire-spread mechanics. Agents must temporarily disengage when they run out of limited suppressant to recharge before rejoining the firefighting efforts.

The MOASEI competition website is here https://oasys-mas.github.io/moasei.html, where details of the competition including competition registration deadline (Apri 1, 2025) and solution submission deadline (April 11, 2025), the available codebase and benchmarks, and rules.

We encourage everyone interested in working in OASYS to participate in this competition!

Sincerely,

Adam Eck, Leen-Kiat Soh, and Prashant Doshi

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cs.umbc.edu/pipermail/agents/attachments/20250304/41f36d1e/attachment.htm>