A Systematic Review of Machine Learning Approaches for Optimizing Production Scheduling in Smart Manufacturing

Abstract

This systematic review investigates the application of machine learning (ML) approaches in optimizing production scheduling within the context of smart manufacturing. As the fourth industrial revolution, Industry 4.0, reshapes manufacturing with interconnected systems, there is a growing demand for innovative solutions to address the inefficiencies in traditional scheduling processes. The impact of various machine learning (ML) techniques, including deep reinforcement learning (DRL), genetic algorithms (GA), particle swarm optimization (PSO), and neural networks (NN), on scheduling flexibility, efficiency, and flexibility in an existing real-time production system is examined in this study. The evaluation lists the primary determinants that impact these ML models' performance, such as algorithm selection, data quality, organizational preparedness, and technology infrastructure. Notwithstanding their great potential, issues with scalability, skilled labor, data quality, and interface with existing systems continue to be obstacles to broader use. The study concludes by summarizing the need for ongoing advancements in machine learning models for infrastructure in order to enable seamless deployment in actual industrial settings and offering manufacturers suggestions on how to optimize production scheduling with AI-based solutions.