[Paper Review] Position Paper: From Multi-Agent Pathfinding to Pipe Routing

Gleb Belov, Liron Cohen+4 more|arXiv (Cornell University)|May 21, 2019

Robotic Path Planning Algorithms14 references4 citations

TL;DR

This paper demonstrates that recent Multi-Agent Pathfinding (MAPF) algorithms can be effectively adapted to solve the 3D Pipe Routing (PR) problem by mapping time-expanded MAPF solutions in x-y-t space to spatial PR solutions in x-y-z space. The key contribution is showing that optimal and bounded-suboptimal MAPF algorithms achieve high solution quality and efficiency on large-scale PR instances, opening new industrial applications for MAPF research.

ABSTRACT

The 2D Multi-Agent Path Finding (MAPF) problem aims at finding collision-free paths for a number of agents, from a set of start locations to a set of goal positions in a known 2D environment. MAPF has been studied in theoretical computer science, robotics, and artificial intelligence over several decades, due to its importance for robot navigation. It is currently experiencing significant scientific progress due to its relevance in automated warehousing (such as those operated by Amazon) and in other contemporary application areas. In this paper, we demonstrate that many recently developed MAPF algorithms apply more broadly than currently believed in the MAPF research community. In particular, we describe the 3D Pipe Routing (PR) problem, which aims at placing collision-free pipes from given start locations to given goal locations in a known 3D environment. The MAPF and PR problems are similar: a solution to a MAPF instance is a set of blocked cells in x-y-t space, while a solution to the corresponding PR instance is a set of blocked cells in x-y-z space. We show how to use this similarity to apply several recently developed MAPF algorithms to the PR problem, and discuss their performance on abstract PR instances. We also discuss further research necessary to tackle real-world pipe-routing instances of interest to industry today. This opens up a new direction of industrial relevance for the MAPF research community.

Motivation & Objective

• To bridge the gap between Multi-Agent Pathfinding (MAPF) and industrial 3D Pipe Routing (PR) by identifying structural similarities.
• To demonstrate that state-of-the-art MAPF algorithms can be applied to PR problems with minimal adaptation.
• To evaluate the performance of MAPF algorithms on abstract PR instances in terms of solution quality, efficiency, and scalability.
• To identify research gaps necessary for tackling real-world pipe-routing problems in industrial settings.

Proposed method

• Maps the 3D Pipe Routing problem into a 2D MAPF problem by treating time as a third spatial dimension, transforming x-y-t space solutions into x-y-z space solutions.
• Uses a discretized 6-neighbor 3D grid graph where each cell is either open or blocked, and pipes are routed as paths between start and goal vertices.
• Applies three MAPF algorithms—Conflict-Based Search (CBS), ECBS(1.01), and ECBS(1.05)—to solve the abstract PR problem with bounded suboptimality guarantees.
• Implements focal search for initial planning and FOCAL-based node selection to improve search efficiency in the PR context.
• Treats each pipe as an agent with fixed start and goal locations, ensuring no two pipes occupy the same cell (vertex) in the solution.
• Employs a cost function based on total path length, minimizing the sum of individual pipe route lengths.

Experimental results

Research questions

• RQ1 Can existing optimal and bounded-suboptimal MAPF algorithms be effectively repurposed for 3D Pipe Routing (PR) problems?
• RQ2 How do MAPF algorithms perform in terms of solution quality, success rate, and computational efficiency on abstract PR instances?
• RQ3 What are the key challenges in adapting MAPF algorithms to real-world pipe-routing scenarios with complex constraints?
• RQ4 How does the structural similarity between MAPF and PR enable cross-domain transfer of algorithmic advances?
• RQ5 What further research is required to make MAPF-based approaches viable for industrial-scale pipe routing applications?

Key findings

• MAPF algorithms, particularly CBS and ECBS variants, successfully solve large-scale abstract PR instances with optimal or near-optimal solutions.
• The ECBS(1.01) and ECBS(1.05) algorithms achieve bounded suboptimality with strong cost guarantees, ensuring high solution quality.
• The performance of MAPF algorithms on PR instances is robust across different environments, with high success rates on problems with hundreds of pipes.
• The time-expanded mapping from x-y-t space to x-y-z space enables direct reuse of advanced MAPF search techniques for spatial routing.
• The results suggest that MAPF research can now target industrial pipe routing as a new, high-impact application domain.
• Existing industrial pipe routing methods are either suboptimal (e.g., prioritized planning) or intractable at scale, making MAPF-based approaches a promising alternative.