Machine Learning and Analytics Applications in Logistics Operations: A Systematic Survey
TL;DR
A systematic review of peer-reviewed articles that apply analytics and machine learning to obtain better logistics operational performance indicates that predictive analytics is the most widely used technique, in demand forecasting and transportation, whereas prescriptive analytics suffers from system integration challenges and computational intensity.
Abstract
The increasing complexity of logistics operations owing to e-commerce development, supply chains, and rising customer expectations has driven organizations to employ decision support systems driven by analytics and machine learning technologies. This study presents a systematic review of peer-reviewed articles that apply analytics and machine learning to obtain better logistics operational performance. Using established systematic review guidelines, papers on analytics and supply chain logistics published between 2007 and 2025 were identified and categorized by logistics functions and analytical techniques. The results indicate that predictive analytics is the most widely used technique, in demand forecasting and transportation, whereas prescriptive analytics suffers from system integration challenges and computational intensity. In addition to improved performance, issues such as data quality, model explainability, and implementation in practice still require further attention. This survey thus brings together research findings and identifies gaps, ultimately guiding future research and analytics-based decision-making in the field of logistics.