GuardPIBT improves large-scale 3D multi-agent path finding efficiency
GuardPIBT: Counterfactually Gated Neural Guidance for Ultra-Large-Scale 3D Multi-Agent Path Finding
Robotics
Summary
Finding paths for many agents moving in crowded 3D spaces is very hard because they can block each other. The authors improved an existing method called PIBT by adding a smart neural network that suggests better movement orders while keeping PIBT’s safety checks. This helps thousands or even 100,000 agents move reliably without getting stuck or causing traffic jams in simulations. Their approach also adapts to very large groups efficiently, making it suitable for complex environments like warehouses.
What this means in practice
- •For warehouse automation teams: Coordinate movement of up to 100,000 robots in complex warehouse layouts with guaranteed collision-free paths.
- •For robotics fleet operators: Manage large fleets of delivery or service robots in crowded 3D environments with improved traffic efficiency and safety.
Authors
Yuan Zhou, Zhenyu Hou, Guangtong Xu, Xiaoqiang Ji, Yuqing Tang, Jialiang Hou, Fei Gao
Abstract
Large-scale 3D multi-agent path finding becomes increasingly difficult under dense traffic. Priority Inheritance with Backtracking (PIBT) scales well, but its one-step goal-directed ordering may become insufficient under dense interactions and large-scale congestion. We present GuardPIBT, which augments rather than replaces the PIBT executor: neural predictions only propose residual reorderings of PIBT's native candidates, while final actions remain determined by PIBT. First, local graph attention models nearby interactions, while global source--goal transport features provide population-level coordination context for candidate reordering. Second, a counterfactual group gate filters reorderings whose closed-loop effects may degrade coordination. Third, for ultra-large populations, population-adaptive grouping preserves decision granularity, asynchronous cached inference amortizes neural computation, and selective repair resolves long-tail agents. PIBT retains validity checking, priority inheritance, and backtracking throughout. Experiments with up to 100,000 agents demonstrate reliable completion across 2D and 3D environments, including all three 100,000-agent warehouse runs with zero audited graph violations. The project website is available at {\color{magenta}\texttt{https://guardpibt.github.io/GuardPIBT/}}.