Lot streaming of hybrid flowshops with variable lot sizes and eligible machines
Document Type
Article
Publication Date
2023
Abstract
Hybrid flowshops are a special type of manufacturing systems, in which a stage may contain identical or unrelated parallel machines. This paper deals with a more practical approach for lot streaming hybrid flowshop in which the sublot sizes of jobs can vary from one stage to the next according to machines’ speed. Two models of mixed-integer nonlinear programming are developed to minimise the make-span of two different hybrid flowshop systems. The first model deals with unrelated parallel machines with eligibility, independent setup time, and variable sublot sizes. The second model is a special case of the hybrid flowshop as it consists of multi-stages comprising one machine at the stages preceding the final stage, while the final stage includes unrelated parallel machines. The first model was studied and the data gathered were analysed using ANOVA test to evaluate the factors’ effect on system. The factors are number of jobs, maximum number of batches, setup time, and machine’s configuration. The analysis revealed that all the factors were effective. The second model was compared to benchmarking published paper and it gets better results.
Recommended Citation
Zaky, E.A., Abdelmaguid, T.F., Mohamed, T.A. and Mohamed, S.T. (2023) ‘Lot streaming of hybrid flowshops with variable lot sizes and eligible machines’, Int. J. Industrial and Systems Engineering, Vol. 43, No. 2, pp.238–264.
Comments
In conclusion, this paper presents a new aspect of lot streaming. It is a system of HFS with unrelated parallel eligible machines and variable sublot sizes to minimise the make-span. The paper introduced two different models of two different systems that can be employed in more practical cases. The first one is HFS with unrelated parallel eligible machines, independent setup time, and variable sublot sizes, while the second system is a special case of HFS as there is one machine at the stages preceding the final stage, whereas the final stage included several unrelated parallel machines. The researchers resorted to the use of mixed-integer nonlinear programming to solve the systems’ problems and minimise the make-span. The authors evaluated the proposed models by using two approaches. As for the first system, the ANOVA test was used to evaluate the performance of the model, and our factors were selected to study its effect on the system. These factors are the number of jobs, the maximum number of batches, machines `configuration, and setup time. The ANOVA test showed that all factors and their interaction were significant and effective. As for the second model, a comparative study was conducted to evaluate its performance, and this proposed second model has a lower Cmax than the benchmarking published paper. Finally, the future work can focus on utilising heuristics to solve large scale real problems.