The Hybrid Approach of a Machine Learning Algorithm for Addressing the Permutation Flow-Shop Scheduling Problem

The existing numerous challenges encompassing raw materials, labour, electricity, machining time, and customer constraints mean that engineering and industrial facilities are under constant pressure to meet demand and increase productivity in the manufacturing sector.Traditional optimisation methods often fall short in effectively addressing these challenges.Recognising the need for advanced solutions, this paper introduces the PFSP as a critical aspect of manufacturing planning, aiming to minimise makespan.Makespan minimisation in PFSP is particularly challenging due to varying processing times and the number of jobs.To tackle this issue, the study advocates for an investigation of machine learning algorithms, emphasising the application of a hybrid approach that integrates the strengths of the Q-learning algorithm and the iterated greedy algorithm (IGA).Hybridized Q-learning along with the IGA algorithm is proposed as a novel modified algorithm.By combining Q-learning with the iterated greedy algorithm, the QIGA enhances its ability to explore the solution space thoroughly.In order to determine the optimal sequence for processing "n" jobs on "m" machines and improve the efficiency of the permutation flow-shop scheduling problem based on makespan, the QIG algorithm was proposed.To evaluate its effectiveness, the algorithm was tested using Taillard benchmark problems.The findings indicate that the suggested algorithm's performance surpasses that of traditional heuristics and metaheuristics approaches and also matches the optimal upper bound makespan value in all instances.Leveraging this effectiveness, two case studies were undertaken, and their outcomes were compared with alternative algorithms.The derived findings and schedules not only contribute to efficient product completion with minimised makespan time but also lead to heightened productivity for manufacturers.QIGA can be integrated with automated manufacturing systems, facilitating real-time scheduling and decision-making.The integration supports Industry 4.0 initiatives and enhances the automation of production processes.