Pengembangan Model Multi-Objective Tata Letak Fasilitas Dinamis Dengan Ketidakpastian Produksi pada Sistem Manufaktur Berbasis Zona

Abstract

Dynamic Facility Layout Planning (DFLP) is related to manufacturing facility planning to be able to respond to changes (changeability) in the form of alignment of departments, zones, workstations and machines by considering dynamic parameters and drivers. There are 3 main gaps according to 7 studies and 3 studies in the literature review that consider environmental dynamics only considering demand and production resources. scheduler dynamics and production strategy execution order; as well as the dynamics of rules such as Safety, management and government regulations and standardization. The dynamic model development carried out by previous research only applies to continuous layouts which are widely applied to chemical process-based industries, while the zone-based manufacturing industry tends to apply its facility layout based on a discrete approach so that this dynamic model is not suitable for application in the manufacturing industry. Previous studies have also developed a mathematical model for zone-based manufacturing industries, but the costs of material handling and re-layout which should be contradictory, in this research are combined so that they do not accommodate the reality of the system that should be. This study develop a DFLP -based dynamic layout model which is a gap in dynamic layout research. The zone based DFLP system is analyzed more comprehensively so as to obtain a model that needs to consider the dynamic aspects of demand, process sequence dynamics and resources. It is necessary to consider dynamic performance measures in the form of changeability parameters. The model was developed in the form of multiple objectives to minimize material handling costs and re-layout costs. The model was tested on 3 cases consisting of dummy data, the aircraft industry from the literature and the car manufacturing industry. Model performance shows that the model produces 100 Pareto optima. This result is said to be good because it has high diversity and does not produce infeasibility with near optimum results. The model is able to provide layout recommendations dynamically with dynamic measures that are more adaptive and flexibility for all cases at the expense of robustness tested by t test with results reject H0 for the car manufacturing industry.