An Improved BiLSTM-Based Method for Master Production Schedule of Medical Equipment

At present, most medical equipment manufacturers still adopt traditional manual methods to compile the master production schedule for orders. To address the issues of low efficiency and poor accuracy in manually generated production schedules, this study proposes a ZOA-BiLSTM-based method for generating master production schedule, with the objective function of minimizing total days required for order production. By training the deep learning model with historical production order data, the method focuses on the correlation between task details, production information, and master production scheduling results, which serve as model inputs. The model's output is the number of days for order production. The Zebra Optimization Algorithm (ZOA) optimizes hyperparameters, such as learning rate, the number of hidden layer nodes, and the number of iterations for the BiLSTM model, which improves the accuracy of the deep learning model. The introduction of the attention mechanism further accelerates the optimization process. Simulation results demonstrate the effectiveness of the ZOA-BiLSTM optimization model in solving master production scheduling problems.