LEARNING OPTIMIZING TECHNIQUES FOR LOGISTICS AND INVENTORY MANAGEMENT

Given that the face of supply chains is constantly rapidly, then the use of ML may potentially create quite several opportunities within this flexible and sustainable realm. This research aims to understand how numerous algorithms within the areas of artificial intelligence particularly, the subfield of machine learning enhance the supply chain together with its inventory systems and flow. For instance, Ridge Regression, Lasso Regression, Elastic Net Regression, Gradient Boosted Trees and K-means Clustering, DBSCAN, and CNNs were designed specifically to solve problems in those fields. The use of both Ridge Regression and Lasso Regression allowed for the accurate improvement of the inventory forecast which in turn decreased forecast error of inventories by 20% and the number of times inventory was turned over by 15%. With the help of GBTs and CNNs the delivery routes were optimized and it contributed to the reduction of the transportation costs by 30% and the delivery time by 25%. By using K-means and DBSCAN, it was easier to sort the inventories and manage risks; which led to an improvement of the inventory turnover ratio by 15 percent and the level of addressing the unanticipated by 18 percent. In addition, the addition of external stimuli, for instance, traffic flow and trending topics on social media boosted the models and defined the differentiation of their decision-making values. They are presumably the sign that with machine learning it is possible to build better, efficient, relevant supply networks. This research sits well with the current literature on the application of ML to the field of supply chain management and offers some practical and useful insights to managers on how to best deploy these techniques for higher organizational value.