Schedule
Time: October 1st: 8:30AM - 12:30AM Eastern Time
Location: Room 321
Time: October 1st: 8:30AM - 12:30AM Eastern Time
Location: Room 321
We define the MAPF problem and walk through the classic algorithms (CBS, PP, PBS, and PIBT).
Stern et al., Multi-Agent Pathfinding: Definitions, Variants, and Benchmarks, SoCS 2019.
Sharon et al., Conflict-Based Search for Optimal Multi-Agent Pathfinding, Artificial Intelligence, 2015.
Ma et al., Searching with Consistent Prioritization for Multi-Agent Path Finding, AAAI 2019.
Okumura et al., Priority Inheritance with Backtracking for Iterative Multi-Agent Path Finding, Artificial Intelligence, 2022.
Sven Koenig, Keisuke Okumura, and Jiaoyang Li
State-lattice MAPF Planning
We show how to plan on a lattice of feasible motions so that the resulting paths respect a robot's kinematics.
Dynamic Bounded MAPF Algorithms
We introduce discontinuity-bounded search, which plans using precomputed motion primitives and enables search-based MAPF to coordinate teams of robots with different dynamics, actuation limits, and shapes.
MAPF as a Protocol for Heterogeneous Fleets of Robots
We present a simple plan() API that takes space-time constraints, so robots with different embodiments and proprietary planners can be coordinated without custom integration work.
Cohen et al., Optimal and Bounded-Suboptimal Multi-Agent Motion Planning, SoCS 2019.
Moldagalieva et al., db-CBS: Discontinuity-Bounded Conflict-Based Search for Multi-Robot Kinodynamic Motion Planning, ICRA 2024.
Moldagalieva et al., db-ECBS: Interaction-Aware Multirobot Kinodynamic Motion Planning, T-RO 2025.
Moldagalieva et al., db-LaCAM: Fast and Scalable Multi-Robot Kinodynamic Motion Planning with Discontinuity-Bounded Search and Lightweight MAPF, ICAPS 2026.
Veerapaneni et al., Conflict-Based Search as a Protocol: A Multi-Agent Motion Planning Protocol for Heterogeneous Agents, Solvers, and Independent Tasks, ICRA 2026.
Keisuke Okumura and Jiaoyang Li
Fast Replanning
We embed lightweight MAPF planners directly into the control loop, replanning fast enough to absorb the uncertainty that makes open-loop execution fail.
Execution with Temporal Plan Graph
We convert a MAPF plan into a dependency graph that preserves the planned ordering, giving collision-free and deadlock-free execution under communication delays and robot breakdowns.
Closing the loop: Learning-Enhanced MAPF Pipeline
We use a learned model of execution time to feed real execution behavior back into planning, so the planner optimizes what actually happens rather than what the simple model predicts.
Shankar et al., LF: Online Multi-Robot Path Planning Meets Optimal Trajectory Control, WoMAPF 2026.
Hönig et al., Persistent and Robust Execution of MAPF Schedules in Warehouses, RA-L 2019.
Feng et al., A Real-Time Rescheduling Algorithm for Multi-Robot Plan Execution, ICAPS 2024.
Su et al., Bidirectional Temporal Plan Graph: Enabling Switchable Passing Orders for More Efficient Multi-Agent Path Finding Plan Execution, AAAI 2024.
Jiang et al., Speedup Techniques for Switchable Temporal Plan Graph Optimization, AAAI 2025
Su et al., BTPG-max: Achieving Local Maximal Bidirectional Pairs for Bidirectional Temporal Plan Graphs, AAAI 2026.
Yan et al., From Discrete Plans to Real-World Execution: A World-Model-Driven Framework for Execution-Aware Multi-Agent Path Finding, arXiv 2026.
Jingkai Chen, Han Zhang
We describe how a production warehouse with hundreds of fast-moving robots is routed, and what had to change in PBS — risk-aware planning with high-fidelity models, fault tolerance, deadlock breaking, and heavy engineering — to meet throughput and runtime targets.
Keisuke Okumura and Jiaoyang Li
Multi-Drone Systems
We show how modeling motion-dependent travel times lets MAPF plans drive aggressive, high-speed quadrotor flight rather than conservative point-to-point hops.
Multi-Robot-Arm Systems
We explain what changes when agents are manipulators: high-dimensional single-agent planning, expensive collision checking, and execution with stochastic actions.
Okumura et al., Concrete multi-agent path planning enabling kinodynamically aggressive maneuvers, npj Robotics 2026.
Shaoul et al. Accelerating Search-Based Planning for Multi-Robot Manipulation by Leveraging Online-Generated Experiences, ICAPS 2024.
Huang et al., APEX-MR: Multi-Robot Asynchronous Planning and Execution for Cooperative Assembly, RSS 2025.
Huang et al., VAMP-MR: Vector-Accelerated Motion Planning and Execution for Multi-Robot-Arms, IROS 2026.
Shaoul et al., Embodying Multi-Hand Manipulation Policies by Searching the Assignment and Null Spaces, SoCS 2026.
LSMART is an open-source, scalable simulator for evaluating lifelong Multi-Agent Path Finding algorithms in realistic warehouse environments. You will run LSMART from a Docker image and compare two planning-window settings on a 150-robot lifelong instance, seeing for yourself how a planning choice shows up in throughput. More details can be found on the LSMART page.
Yan et al., Advancing MAPF Toward the Real World: A Scalable Multi-Agent Realistic Testbed (SMART), RA-L 2026.
Yan et al., Lifelong Scalable Multi-Agent Realistic Testbed and A Comprehensive Study on Design Choices in Lifelong AGV Fleet Management Systems, arXiv 2026.