Metamodeling and multi-objective optimization for buffer allocation in multi-product serial production lines
Ingeniería Investigación y Tecnología
Resumen
The allocation of capacity for the work in process is a complex optimization problem due to the stochastic nature of queueing systems; in addition, the variant with several classes of jobs has been scarcely addressed and so, there is a lack of proposals to quantify the buffer size in such situation. In this work we characterize the behavior of the throughput and cycle time as a function of buffer sizes in a serial production line with reliable stations and more than one job class. A simulation model built in ARENA software R16, a surface experimental design and regression analysis were used to obtain the approximate expressions of the throughput and the cycle time as a function of buffer size. The buffer sizes were obtained using a multi-objective optimization approach. The results show that, when all work classes are in the system, to maximize throughput, a greater buffer size must be allocated to slow stations; to minimize cycle time, the best allocation is a small buffer size at the fastest stations; the capacity allocated to each buffer is an indicator of slow and fast stations; finally the actual service time at a station is the result of the weighted service time of each work class. The proposal is replicable and flexible enough to add new variables; it will be useful for engineers and administrators who must plan the buffer size in systems that serve two or more work classes with random service times.
Datos de publicación
Cita esta publicación
Comprueba el estilo que requiere tu institución o revista.
Fuentes y trazabilidad
Se conserva la procedencia de los datos y se da prioridad a la revista cuando existe información editorial.
Licencias declaradas por las fuentes
https://creativecommons.org/licenses/by-nc/4.0/