Sahar Tadayonirad, Hany Seidgar, Hamed Fazlollahtabar and Rasoul Shafaei
In real manufacturing systems, schedules are often disrupted with uncertainty factors such as random machine breakdown, random process time, random job arrivals or job…
Abstract
Purpose
In real manufacturing systems, schedules are often disrupted with uncertainty factors such as random machine breakdown, random process time, random job arrivals or job cancellations. This paper aims to investigate robust scheduling for a two-stage assembly flow shop scheduling with random machine breakdowns and considers two objectives makespan and robustness simultaneously.
Design/methodology/approach
Owing to its structural and algorithmic complexity, the authors proposed imperialist competitive algorithm (ICA), genetic algorithm (GA) and hybridized with simulation techniques for handling these complexities. For better efficiency of the proposed algorithms, the authors used artificial neural network (ANN) to predict the parameters of the proposed algorithms in uncertain condition. Also Taguchi method is applied for analyzing the effect of the parameters of the problem on each other and quality of solutions.
Findings
Finally, experimental study and analysis of variance (ANOVA) is done to investigate the effect of different proposed measures on the performance of the obtained results. ANOVA's results indicate the job and weight of makespan factors have a significant impact on the robustness of the proposed meta-heuristics algorithms. Also, it is obvious that the most effective parameter on the robustness for GA and ICA is job.
Originality/value
Robustness is calculated by the expected value of the relative difference between the deterministic and actual makespan.