AlphaTruss: Monte Carlo Tree Search for Optimal Truss Layout Design

Truss layout optimization under complex constraints has been a hot and challenging problem for decades that aims to find the optimal node locations, connection topology between nodes, cross-sectional areas of connecting bars. Monte Carlo Tree Search (MCTS) is reinforcement learning search technique competent solve decision-making problems. Inspired by success AlphaGo using MCTS, truss formulated as Markov Decision Process (MDP) model, 2-stage MCTS-based algorithm, AlphaTruss, proposed generating considering topology, geometry, bar size. In this MDP three sequential action sets adding bars, selecting sectional greatly expand solution space reward function gives feedback actions according both geometric stability structural simulation. To actions, AlphaTruss solves model best decision in each design step searching through MCTS. Compared with existing results from literature, exhibits better performance finding minimum weight stress, displacement, buckling constraints, which verifies validity efficiency established algorithm.