Ensemble Machine Learning and Knowledge-Based Genetic Algorithm for Energy-Efficient Scheduling and an Empirical Study

Overall energy efficiency is crucial for green production and carbon dioxide emission, especially for energy-intensive industries such as steel making. To deal with unrelated parallel machine scheduling for steelmaking-continuous casting that is the bottleneck and energy-intensive process, this study aims to develop an effective approach that integrates ensemble machine learning and hybrid genetic algorithm for energy-efficient scheduling and cleaner production. The knowledge-based method is used to train intelligent job dispatching agent for providing good quality initial scheduling solutions. The hybrid genetic algorithm is developed to obtain the optimal scheduling solution to address dynamic multiple objective production scheduling problem. To estimate the validity of the proposed approach, experiments are designed in different scenarios based on realistic data of a leading steelmaking company in Asia. Experimental results have shown that the proposed approach can generate high productivity and energy efficient solution for the present problem efficiently and effectively