Objectives  The disorder of different types of slotting equipment occurs during the use of an automated 3D warehouse, affecting the efficiency of subsequent material support. Therefore, it is necessary to carry out research on the slotting optimization of automated 3D warehouses.

  Methods  First, a slotting optimization model is established to minimize the inventory time and distance of the same group of goods after analyzing the operation characteristics of the automated 3D warehouse. Next, in order to overcome the shortcomings of the traditional Monte Carlo Tree Search (MCTS) algorithm that tends to fall into local optimality, a simulated annealing algorithm is integrated with the MCTS algorithm to optimize the node selection operation. Finally, the improved MCTS algorithm is tested for its optimization, stability and convergence.

  Results  The improved MCTS algorithm is at least 30% more optimized than the greedy, Rubik's cube reduction and traditional MCTS algorithms in terms of slotting optimization running time.

  Conclusions  By adding the optimal path random selection factor to the MCTS algorithm, it can be prevented from falling into local optimality, effectively achieving slotting optimization.