CTS-CBS: A New Approach for Multi-Agent Collaborative Task Sequencing and Path Finding

2025-03-26

AI Summary

This paper addresses the Collaborative Task Sequencing–Multi-Agent Path Finding (CTS-MAPF) problem, where multiple agents must visit a sequence of intermediate task locations in a globally prescribed order before reaching individual goals, all while avoiding collisions. We propose the first two-level decoupled framework: (i) a high-level layer models task sequencing as a joint Traveling Salesman Problem (jTSP) and searches over task-sequence forests; (ii) a low-level layer extends Conflict-Based Search (CBS) to generate time-ordered, collision-free single-agent paths under temporal constraints. We prove theoretical completeness and ω-suboptimality, enabling flexible quality–efficiency trade-offs. Evaluated on CTS-MAPF and MG-MAPF benchmarks, our method achieves a 20× improvement in success rate, 100× speedup in solving time, and incurs <10% path cost overhead versus optimal baselines. Real-robot experiments further validate deployment feasibility.

Technology Category

• Multiagent Systems: Multiagent Planning
• Planning, Routing, and Scheduling: Temporal Planning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

This paper addresses a generalization problem of Multi-Agent Pathfinding (MAPF), called Collaborative Task Sequencing - Multi-Agent Pathfinding (CTS-MAPF), where agents must plan collision-free paths and visit a series of intermediate task locations in a specific order before reaching their final destinations. To address this problem, we propose a new approach, Collaborative Task Sequencing - Conflict-Based Search (CTS-CBS), which conducts a two-level search. In the high level, it generates a search forest, where each tree corresponds to a joint task sequence derived from the jTSP solution. In the low level, CTS-CBS performs constrained single-agent path planning to generate paths for each agent while adhering to high-level constraints. We also provide heoretical guarantees of its completeness and optimality (or sub-optimality with a bounded parameter). To evaluate the performance of CTS-CBS, we create two datasets, CTS-MAPF and MG-MAPF, and conduct comprehensive experiments. The results show that CTS-CBS adaptations for MG-MAPF outperform baseline algorithms in terms of success rate (up to 20 times larger) and runtime (up to 100 times faster), with less than a 10% sacrifice in solution quality. Furthermore, CTS-CBS offers flexibility by allowing users to adjust the sub-optimality bound omega to balance between solution quality and efficiency. Finally, practical robot tests demonstrate the algorithm's applicability in real-world scenarios.

Problem

• Multi-agent pathfinding with ordered task sequencing
• Collision-free path planning for collaborative agents
• Balancing solution optimality and computational efficiency

Innovation

• Two-level search for task sequencing and pathfinding
• Generates joint task sequences from jTSP solution
• Adjustable sub-optimality bound for efficiency balance