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

Abstract

Demand forecasting is a quite challenging task, which is sensitive to several factors such as endogenous and exogenous parameters. In the context of supply chain management, demand forecasting aids to optimize the resources effectively. In recent years, numerous methods were developed for Supply Chain (SC) demand forecasting, which posed several limitations related to inadequate handling of dynamic time series patterns and data requirement problems. Thus, this research proposes a Migrative Armadillo Optimization-enabled one-dimensional Quantum convolutional neural network (MiA + 1D-QNN) for effective demand forecasting. The Migrative Armadillo Optimization (MAO) algorithm effectively optimizes the hyperparameters of the model. Specifically, the 1D-QNN approach offers exponential speed in the forecasting tasks as well as provides accurate prediction. Furthermore, the K-nearest Neighbor imputation technique fills the missing values, which preserves the data integrity as well as reliability. The experimental outcomes attained using the proposed model achieved a correlation of 0.929, Mean Square Error (MSE) of 7.34, Mean Absolute Error of 1.76, and Root Mean Square Error (RMSE) of 2.71 for the supply chain analysis dataset. For DataCo smart SC for big data analysis dataset, the MiA + 1D-QNN model achieved the correlation of 0.957, Mean Square Error (MSE) of 6.00, Mean Absolute Error of 1.62, and Root Mean Square Error (RMSE) of 2.45.