Optimisation of complex and large-sized single-row facility layout problems with a unique hybrid meta-heuristic framework (Q2627358)

Summary: This paper proposes a hybrid framework based on genetic algorithms (GAs) and discrete-event simulation (DES) for optimisation of large-sized single-row facility layout problems (SRFLPs). For a SRFLP with \(n\) number of facilities, there are \(n!\) layout formations that have to be modelled and evaluated so as to find the optimal single-row formation of facilities. For that reason, as the number of facilities gets larger, the solution space grows exponentially and so evaluation of all potential layout formations could be extremely hard or even impossible. For dealing with non-deterministic polynomial-time hardness of large-sized SRFLPs, this study puts forward a novel GA wherein DES is applied as a performance evaluation tool for calculating the fitness function. The proposed framework employs DES for modelling and evaluating diverse layout formations. Manufacturing lead-time is taken into consideration as the performance evaluation measure. The solution quality is investigated through a real case study in an injection moulding process with sequence-dependent setup times in a refrigerator manufacturing company. Notably, significance of the proposed framework in comparison with previous studies in this area lies in integrating GAs and DES for optimising large-sized SRFLPs in presence of complexity, non-linearity and stochasticity.