Published Online:May 2026
Product Name:The IUP Journal of Operations Management
Product Type:Article
Product Code:IJOM020526
DOI:10.71329/IUPJOM/2026.25.2.28-47
Author Name:Yogesh Babu K and Anil B Gowda
Availability:YES
Subject/Domain:Management
Download Format:PDF
Pages:28-47
The study develops a hybrid decision-support framework that integrates Machine Learning (ML) and Deep Learning (DL) models with a Binary Linear Programming (BLP) optimization approach for predictive and risk-informed supplier selection. The framework initially utilizes supervised learning techniques to forecast supplier delivery delays and assess quality-related risks based on historical manufacturing data. A continuous training mechanism is embedded within the predictive architecture to enable model adaptation as new operational data becomes available, thereby enhancing predictive accuracy and robustness over time. The resulting risk estimates are subsequently incorporated into a binary optimization model to determine optimal supplier selection and multiperiod order allocation decisions, while minimizing total cost and exposure to disruptions. In addition to improving operational performance, the proposed framework contributes to sustainability objectives by reducing waste arising from delayed deliveries, lowering reliance on emergency logistics, minimizing excess inventory buffers, and mitigating carbon-intensive disruptions.
Manufacturing supply chains have become increasingly complex due to globalization, shortened product life cycles, and a growing dependence on multi-tier supplier networks. While these developments have enabled cost-efficiencies and access to specialized capabilities, they have also increased exposure to supplier-related risks such as delivery delays, quality failures, and inconsistent response times.