Multi-objective dynamic job shop scheduling optimization in manufacturing systems: A short review

The growing emphasis on accountability and sustainability, along with the recent push to manage sudden disruptions, has added new layers of complexity to modern Job Shop Manufacturing Systems (JSMs). Even small reductions in conflicting objectives can have ripple effects throughout these systems. This review brings together current research on the use of Multi-Objective Optimization (MOO) methods in dynamic scheduling, with particular attention to how energy cost optimization (ECO) is incorporated into real-time decision-making frameworks. A critical review of current methodologies reveals that achieving sustainability necessitates balancing the inherent time versus cost/energy trade-off, driving the adoption of metaheuristics like Genetic Algorithms (GA) and advanced Deep Reinforcement Learning (DRL) for adaptive policy generation. We detail the indispensable role of integrated digital architectures, including Digital Twin (DT) frameworks for virtual validation and the Internet of Things (IoT) for real-time data acquisition. This synergy facilitates the shift from static planning to autonomous, resilient control, demonstrating proven effectiveness in reducing total tardiness, carbon emissions, and operational costs. Finally, outlining key challenges particularly computational scalability and model generalization and proposing future research directions focused on hybrid optimization, edge computing, and comprehensive Industry 5.0 integration for truly sustainable manufacturing concludes the paper.