Mastering truss structure optimization with tree search

TL;DR

This paper presents a novel approach combining generative grammar and Monte Carlo Tree Search for efficient and robust optimization of truss structures, outperforming traditional reinforcement learning methods.

Contribution

It introduces a new MCTS-based framework that effectively handles large state spaces and intermediate construction stages in truss design, demonstrating superior performance over existing RL techniques.

Findings

MCTS outperforms Q-learning and deep Q-learning in robustness and efficiency

The approach effectively navigates large, complex state spaces

Early decision nodes are crucial for high-quality solutions

Abstract

This study investigates the combined use of generative grammar rules and Monte Carlo Tree Search (MCTS) for optimizing truss structures. Our approach accommodates intermediate construction stages characteristic of progressive construction settings. We demonstrate the significant robustness and computational efficiency of our approach compared to alternative reinforcement learning frameworks from previous research activities, such as Q-learning or deep Q-learning. These advantages stem from the ability of MCTS to strategically navigate large state spaces, leveraging the upper confidence bound for trees formula to effectively balance exploitation-exploration trade-offs. We also emphasize the importance of early decision nodes in the search tree, reflecting design choices crucial for highly performative solutions. Additionally, we show how MCTS dynamically adapts to complex and extensive state spaces without significantly affecting solution quality. While the focus of this paper is on truss optimization, our findings suggest MCTS as a powerful tool for addressing other increasingly complex engineering applications.