Publication

H. He, X. Wu, S. Zhao and Z. Ren. Conflict-Based Search for Multi-Agent Path Finding with Elevators. In IEEE International Conference on Automation Science and Engineering (CASE), 2026.

Abstract: This paper studies Multi-Agent Path Finding with Elevators (MAPF-E), where agents may need to move across floors by using elevators. The elevator actions introduce long-duration interactions and require the planner to account for elevator states in addition to agent locations. The paper proposes CBS-E, an optimal Conflict-Based Search extension with elevator-specific constraints, and MDD-E, an elevator-aware multi-valued decision diagram method for choosing conflicts. The experiments show substantially higher success rates than the baseline and large reductions in the number of CBS-E iterations.

Download the paper in pdf.