Hybrid Quantum-Classical Multi-Agent Pathfinding

Thore Gerlach, Loong Kuan Lee, Frederic BARBARESCO, Nico Piatkowski

TL;DR: First optimal quantum algorithms for Multi-Agent Pathfinding based on generating new paths until an optimality criterium is met.

Abstract: Multi-Agent Path Finding (MAPF) focuses on determining conflict-free paths for multiple agents navigating through a shared space to reach specified goal locations. This problem becomes computationally challenging, particularly when handling large numbers of agents, as frequently encountered in practical applications like coordinating autonomous vehicles. Quantum Computing (QC) is a promising candidate in overcoming such limits. However, current quantum hardware is still in its infancy and thus limited in terms of computing power and error robustness. In this work, we present the first optimal hybrid quantum-classical MAPF algorithms which are based on branch-and-cut-and-prize. QC is integrated by iteratively solving QUBO problems, based on conflict graphs. Experiments on actual quantum hardware and results on benchmark data suggest that our approach dominates previous QUBO formulations and state-of-the-art MAPF solvers.

Lay Summary: Coordinating multiple agents—like drones or delivery robots—to move without collisions is a complex and critical challenge for urban logistics and automated warehouses. This paper introduces two new hybrid quantum-classical algorithms, to solve these large-scale pathfinding problems optimally. By combining classical computing with quantum techniques, the algorithms break the problem into smaller tasks using a mathematical method called Quadratic Unconstrained Binary Optimization (QUBO), which quantum computers handle well. They start with initial paths for each agent and refine them iteratively, using quantum computing to identify better routes while checking for conflicts with classical methods. The algorithms repeatedly adjust and optimize until a collision-free, efficient set of paths is found. The developed algorithms are tested using both classical computers and actual quantum hardware, showing that the proposed hybird method can outperform previous approaches even with current, limited quantum devices. This work highlights how blending quantum and classical computing can already improve real-world applications, such as managing urban drone deliveries or robot fleets. As quantum hardware continues to advance, this hybrid approach promises even greater efficiency and scalability.

Primary Area: Optimization->Discrete and Combinatorial Optimization

Keywords: QUBO, MAPF, Column Generation, Optimal, Quantum Computing, ILP, Operations Research

Submission Number: 8141