Attention-Based Hierarchical Collaborative Optimization for Multi-Agent Path Planning and Resource Allocation

View at publisher Summarise with VIRAT
Authors: Wanxuan Gao, Yingqi Zhu, Ruimeng Wang
Cited by: 0
References: 1

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Path length and battery amp-hours constrain each other. Ignoring this coupling produces deadlocks in warehouses and mid-air returns for drones. We fold the two sub-problems into one two-layer model. The top layer runs self-attention only on agent pairs whose swept volumes overlap within the next At; the bottom layer casts every future energy or CPU request as a query vector and matches it against the closest, fastest, least loaded supplier. Information is exchanged every 0.5 s, so a reroute immediately triggers a re-allocation and vice-versa. Tests with fifty AGVs and twenty UAVs show completion time down 18–25 % and resource utilisation up 13–19 % when the fleet size swings from 10 to 50. The code is public.

Fields of Study & Subjects

Robotic Path Planning Algorithms Advanced Manufacturing and Logistics Optimization Vehicle Routing Optimization Methods

Publication details

Journal / source
IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
Pages
1232–1236
Publication date
28 January 2026
Language
EN