Intelligent Classification and Sorting Optimization of Transported Items Combined with BiLSTM Model

In view of the low efficiency and large error of the traditional method of transport item classification and sorting, this paper proposes an optimization method based on the bidirectional long short-term memory network (BiLSTM) model. First, the multi-dimensional feature data of the items, including size, weight, shape, material, and historical location information, are collected to construct a comprehensive feature vector. Then, the BiLSTM model is used to train the time series data, and the time dependency is captured through forward and backward propagation. Finally, the proximal policy optimization (PPO) strategy is combined to optimize the transportation path and sequence to improve the system adaptability. This method can significantly reduce the transportation time (such as from 38.87 minutes to 18.58 minutes in Experiment 1) and the path error (from 7.16 meters to 1.13 meters in Experiment 3). This study provides an efficient and reliable optimization solution for intelligent warehousing systems.