Migrative armadillo optimization enabled a one-dimensional quantum convolutional neural network for supply chain...

Area of Science:

• Supply Chain Management
• Artificial Intelligence
• Data Science

Background:

• Demand forecasting is critical for supply chain optimization but faces challenges with dynamic time series and data requirements.
• Existing methods often struggle with complex patterns and data integrity issues.
• Accurate demand forecasting improves resource allocation and operational efficiency.

Purpose of the Study:

• To propose an advanced demand forecasting model for supply chain management.
• To address limitations of existing methods in handling dynamic time series and data quality.
• To enhance the accuracy and reliability of demand predictions.

Main Methods:

• Developed a Migrative Armadillo Optimization-enabled one-dimensional Quantum Convolutional Neural Network (MiA + 1D-QNN).
• Utilized Migrative Armadillo Optimization (MAO) for hyperparameter tuning.
• Employed K-nearest Neighbor imputation for handling missing data values.

Main Results:

• The MiA + 1D-QNN model achieved a correlation of 0.929 and MSE of 7.34 on a supply chain analysis dataset.
• On the DataCo smart SC dataset, the model reached a correlation of 0.957 and MSE of 6.00.
• The model demonstrated high accuracy with low Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) values.

Conclusions:

• The proposed MiA + 1D-QNN model offers a significant advancement in demand forecasting for supply chains.
• The integration of MAO and 1D-QNN provides efficient and accurate predictions.
• The method effectively handles data complexities, improving overall supply chain resource management.