Operational Analytics Systems For Real-Time Data-Driven Optimization

Abstract

Supply chain management faces challenges from dynamic demand, disruptions, and high-dimensional data. Traditional static batch methods lack responsiveness in changing environments. This research proposes a smart dynamic operational analytics system for optimized, data-driven supply chain decision-making. The architecture proposed integrates data streams using Smart Logistics Supply Chain Dataset with 1000 records. Principal Component Analysis (PCA) is used for feature extraction, dimensionality reduction, and consistency Min-Max normalization in data preparation. Processed features are given to the Wingsuit Flying Search driven Intelligent Backpropagation Neural Network (WFS-IntBPNN) for predicting the state of the system. Intelligent weights are optimized by WFS and learning and prediction of demands, anomalies, and system states are completed by IntBPNN. Experimental results are obtained by Python (version 3.10) and The data obtained indicate that the proposed WFS-IntBPNN model performs superior to the current models, which gives the minimum error value (MSE = 1.8420, MAE = 0.9820). In sum, the system is useful for optimizing resources and accelerating decision-making, for more accurate forecasting, faster response times and higher throughput in high-volume, high-variability situations.