.png){kind=icon}
Return
Multi-agent task assignment and path finding for heterogeneous AGV systems
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2025.130973.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Heterogeneous automated guided vehicle (AGV) fleets in modern warehouses and workshops face coupled challenges, such as capability-aware task–vehicle matching, real-time arrivals, shared resource contention, and deadlock-prone multi-agent path finding (MAPF). This study develops an integrated framework that unifies rolling-horizon task assignment with conflict-aware routing. Task assignment enforces a capability compatibility matrix and optimizes a composite objective that shortens inter-task connection distance while balancing workload and goal-point counts. Routing builds on enhanced conflict-based search (ECBS) with duplicate-destination preprocessing, horizon truncation for synchronization, and a proactive avoidance-point mechanism to prevent and clear deadlocks. Experiments in three industrial scenarios show that the fastest convergence for assignment can achieve 40 s for 490 tasks per hour. Moreover, ECBS maintains short planning times with low deadlock incidence. Based on that, the bounded-suboptimal search paired with explicit deadlock handling scales effectively to dense, heterogeneous fleets, and fast-converging metaheuristics are preferable for real-time assignment in bursty task streams.
Journal
IF:
7.5
Papers:
3.0W
Citations:
10.2W
Organization
Cited Papers
A Comprehensive Review on Leveraging Machine Learning for Multi-Agent Path Finding.png)
IEEE ACCESS
OAAI
IEEE ACCESSIF3.6
OAAIShareSaveShareSaveShareSave
ShareSaveShareSaveShareSaveShareSaveZebra Optimization Algorithm: A New Bio-Inspired Optimization Algorithm for Solving Optimization AlgorithmIEEE ACCESSOAAI
IEEE ACCESSIF3.6
OAAIShareSave
View more