Stepping into Industry 4.0-based optimization model: a hybrid of the NSGA-III and MOAOA
Yaser Sadati-Keneti,Mohammad Vahid Sebt,Reza Tavakkoli-Moghaddam,Armand Baboli,Misagh Rahbari
DOI: https://doi.org/10.1108/k-08-2023-1580
IF: 2.352
2024-06-15
Kybernetes
Abstract:Purpose Although the previous generations of the Industrial Revolution have brought many advantages to human life, scientists have been looking for a substantial breakthrough in creating technologies that can improve the quality of human life. Nowadays, we can make our factories smarter using new concepts and tools like real-time self-optimization. This study aims to take a step towards implementing key features of smart manufacturing including preventive self-maintenance, self-scheduling and real-time decision-making. Design/methodology/approach A new bi-objective mathematical model based on Industry 4.0 to schedule received customer orders, which minimizes both the total earliness and tardiness of orders and the probability of machine failure in smart manufacturing, was presented. Moreover, four meta-heuristics, namely, the multi-objective Archimedes optimization algorithm (MOAOA), NSGA-III, multi-objective simulated annealing (MOSA) and hybrid multi-objective Archimedes optimization algorithm and non-dominated sorting genetic algorithm-III (HMOAOANSGA-III) were implemented to solve the problem. To compare the performance of meta-heuristics, some examples and metrics were presumed and solved by using the algorithms, and the performance and validation of meta-heuristics were analyzed. Findings The results of the procedure and a mathematical model based on Industry 4.0 policies showed that a machine performed the self-optimizing process of production scheduling and followed a preventive self-maintenance policy in real-time situations. The results of TOPSIS showed that the performances of the HMOAOANSGA-III were better in most problems. Moreover, the performance of the MOSA outweighed the performance of the MOAOA, NSGA-III and HMOAOANSGA-III if we only considered the computational times of algorithms. However, the convergence of solutions associated with the MOAOA and HMOAOANSGA-III was better than those of the NSGA-III and MOSA. Originality/value In this study, a scheduling model considering a kind of Industry 4.0 policy was defined, and a novel approach was presented, thereby performing the preventive self-maintenance and self-scheduling by every single machine. This new approach was introduced to integrate the order scheduling system using a real-time decision-making method. A new multi-objective meta-heuristic algorithm, namely, HMOAOANSGA-III, was proposed. Moreover, the crowding-distance-quality-based approach was presented to identify the best solution from the frontier, and in addition to improving the crowding-distance approach, the quality of the solutions was also considered.
computer science, cybernetics