RL-Lock: Reinforcement Learning for Generating Interlocking Assemblies

2026-08-03

AI Summary

This work addresses the problem of automatically decomposing a given 3D voxelized object into a specified number of interlocking parts that can be stably assembled without adhesives or fasteners, relying solely on geometric constraints. To this end, it formulates interlocking assembly generation as a sequential decision-making problem for the first time and introduces a reinforcement learning framework that integrates structured action decomposition with Monte Carlo Tree Search (MCTS)-guided policy-value learning, thereby eliminating reliance on traditional handcrafted heuristics. The proposed method efficiently produces high-quality interlocking structures even for complex and challenging instances, significantly outperforming existing approaches, which often fail or suffer from poor efficiency due to combinatorial explosion in the search space.

Abstract

An interlocking assembly is an assembly in which component parts are connected purely through their geometric arrangement, without relying on external connectors such as glue and nails. Such assemblies have been widely used in a variety of real-world applications due to their structural stability. The problem of generating interlocking assemblies is generally formulated as a shape decomposition problem, where a target 3D object represented as a voxel grid is partitioned into a prescribed number of interlocking pieces. We observe that generating interlocking assemblies is inherently a sequential decision-making problem, where an agent repeatedly decides which piece each voxel should be assigned to. Inspired by the observation, we propose the first reinforcement learning framework RL-Lock for generating interlocking assemblies, without relying on handcrafted search heuristics as existing works did. RL-Lock combines structured action chunking with MCTS-guided policy-value learning to efficiently navigate the large combinatorial search space for interlocking assembly generation. We demonstrate through experiments that RL-Lock allows effective generation of interlocking assemblies, especially for challenging cases in which existing approaches take too long or even fail to find a valid solution.

Problem

• interlocking assemblies
• shape decomposition
• 3D object
• geometric arrangement
• voxel grid

Innovation

• reinforcement learning
• interlocking assemblies
• structured action chunking
• MCTS-guided policy-value learning
• shape decomposition