Research map / Multi-agent systems and games / Cooperative multi-agent RL
Scalable Multi-agent Covering Option Discovery based on Kronecker Graphs
Jiayu Chen, Jingdi Chen, Tian Lan, Vaneet Aggarwal
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use multi-agent covering options instead of single-agent options or no options | MARL agents on multi-agent Grid Maze and Mujoco tasks · Multi-agent goal-achieving tasks | time steps to complete the task | Improved |
| Use four-agent covering options | Distributed Q-learning · 4-agent Grid Maze | task completion: ~400 episodes | Improved |
| Use pairwise options rather than no options or single-agent options | MARL agents · m × n Grid Maze | learning speed: about two orders of magnitude | Improved |
| Use pairwise options with Distributed Q-Learning | MARL agents · 6-agent Grid Maze with random grouping | task completion | Hurt |
| Use pairwise options with Centralized Q-Learning + Force instead of Distributed Q-Learning | MARL agents using pairwise options · 6-agent Grid Maze with random grouping | task completion | Improved |
| Use Centralized Q-Learning + Force rather than Distributed Q-Learning to select options | MARL agents using multi-agent options · Grid Maze tasks | convergence speed and final performance | Improved |
| Approximate the joint transition graph with a Kronecker product of individual-agent graphs | Multi-agent option discovery · Grid Room approximation study, ρ=0.5 | Fiedler estimation accuracy: 100 | No clear effect |
| Approximate the joint transition graph with a Kronecker product of individual-agent graphs | Multi-agent option discovery · Grid Room approximation study, ρ=0.2 | Fiedler estimation accuracy: 99.9; 96.2; 89.8; 79.4 | Mixed |
| Increase transition influence between agents | Kronecker graph approximation · Grid Room approximation study | Fiedler estimation accuracy, ρ=0.2: 99.9; 96.2; 89.8; 79.4 | Hurt |
| Use multi-agent options despite dynamic transition influence among agents | Centralized Q-Learning + Force · Grid Room with random grouping and collisions | task completion steps: 60-100 steps | Improved |
Closest papers by meaning
- Inter-Agent Relative Representations for Multi-Agent Option Discovery (ICLR 2026 · similarity 0.91)
- Option Discovery in the Absence of Rewards with Manifold Analysis (ICML 2020 · similarity 0.86)
- MALinZero: Efficient Low-Dimensional Search for Mastering Complex Multi-Agent Planning (NeurIPS 2025 · similarity 0.85)
- Graph Diffusion for Robust Multi-Agent Coordination (ICML 2025 · similarity 0.84)
- Finite-Time Frequentist Regret Bounds of Multi-Agent Thompson Sampling on Sparse Hypergraphs (AAAI 2024 · similarity 0.84)
- Decentralized Monte Carlo Tree Search for Partially Observable Multi-Agent Pathfinding (AAAI 2024 · similarity 0.84)
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.
