An efficient simulation–neural network–genetic algorithm for flexible flow shops with sequence-dependent setup times, job deterioration and learning effects

• Published in: Neural computing & applications Vol. 31; no. 9; pp. 5327 - 5341
• Main Authors: Azadeh, A., Goodarzi, A. Hasani, Kolaee, M. Hasannia, Jebreili, S.
• Format: Journal Article
• Language: English
• Published: London Springer London 01.09.2019; Springer Nature B.V
• Subjects: Artificial Intelligence; Artificial neural networks; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Computer simulation; Data Mining and Knowledge Discovery; Deterioration; Genetic algorithms; Image Processing and Computer Vision; Integrated approach; Learning theory; Neural networks; Optimization; Original Article; Probability and Statistics in Computer Science; Production scheduling; Sequences; Sequential scheduling; Setup times; Solution space; Learning effects; Deterioration of jobs; Artificial neural network; Genetic algorithm; Dependent setup times; Flexible flow shop
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-018-3368-6

This study presents an integrated approach based on artificial neural network (ANN), genetic algorithm (GA) and computer simulation to explore all the solution space in stochastic flexible flow shop with sequence-dependent setup times, job deterioration and learning effects. The objective of this study is minimizing total tardiness of jobs in the sequences. In this study, the outputs of ANN are inputted to GA and outputs of simulation model are inputted to ANN. We consider learning effects in this problem which means that workers become more experienced with the passage of time, and thus, the processing duration decreases. Deterioration of job means that processing time is a decreasing function of its execution start time. It is not possible to propose a mathematical optimization model for the stated problem; therefore, a simulation optimization approach based on ANN–GA is introduced for a relatively large problem. Finally, actual experiments are conducted to show the applicability of the proposed novel algorithm in finding near-optimal solutions with normal, uniform and exponential processing and setup times. This is the first study that presents an integrated intelligent approach for optimal solution of stochastic flexible flow shop problem with sequence-dependent setup times, job deterioration and learning effects in a real case study.