Enhancing digital manufacturing efficiency and dominance relation driven big Data analytics

Abstract

The integration of AI technologies based on large language models (AI-LLM) with big data analytics has revolutionized digital manufacturing and enabled real-time decision-making in operations and supply chain management (OSCM). However, traditional data envelopment analysis (DEA) models face prohibitive computational complexity in large-scale environments, hindering their adoption for such AI-LLM-driven optimization tasks as demand forecasting, inventory control, and energy efficiency enhancement. To bridge this gap, we propose a dominance–relation-driven DEA framework tailored for big data environments contexts in digital manufacturing. Our approach leverages spatial relationship characteristics and grouping algorithms to reduce computational complexity by 10–30 times, as validated through numerical simulations on industrial datasets. A case study on the Chaohu Lake watershed further demonstrates its practical value in LLM-enhanced environmental monitoring and sustainable supply chain design. This research presents a scalable solution for optimizing efficiency in digital manufacturing, addressing critical challenges in predictive analytics and resource allocation.