Goal programming and multiple criteria data envelopment analysis combined with optimization and Monte Carlo simulation: An application in railway components. Issue 2 (28th September 2021)

• Record Type: Journal Article
• Authors: da Silva, Aneirson Francisco; Silva Marins, Fernando Augusto; Dias, Erica Ximenes; de Carvalho Miranda, Rafael
• Other Names: Wu Desheng Dash guestEditor.; Hall Jon guestEditor.; Belezamo Baloka guestEditor.; Eken Süleyman guestEditor.; Avci Cafer guestEditor.
• Is Part Of: Expert systems. Volume 39:Issue 2(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-09-28
• Subjects: efficiency -- goal programming -- industrial processes -- multiple criteria data envelopment analysis -- optimization via Monte Carlo simulation -- railway and industrial castings; Expert systems (Computer science); 006.33
• DOI: 10.1111/exsy.12840
• Languages: English
• ISSNs: 0266-4720

Abstract

Abstract: This work has been developed in a large steel industry in Brazil, which produces railway and industrial components, and whose aim was to reduce casting defects. Usually, in industrial processes, identifying the causes of defects and their control are relatively complex activities, due to the many variables involved. In this context, the production processes of seven products, involving 38 process variables (inputs and outputs), have been evaluated adopting a new and innovative procedure. Initially, using a Weighted Goal Programming ‐ Multiple Criteria Data Envelopment Analysis (WGP‐MCDEA) model, we identified the most relevant input and output variables, and the studied company validated the results. Next, using the multiple regression technique, empirical functions were constructed for two response variables chosen by the company – number of external cracks and number of internal cracks. Then, to model the real processes adequately, we introduced the occurrence of uncertainty on the coefficients of these functions, considering them as random variables, according to triangular probability functions. Finally, applying the optimizer Optquest, optimization via Monte Carlo simulation (OvMCS) was performed, and with the Ordinary Least Square technique, we obtained the best fit for the two response variables. Specialists from the company validated the proposed procedure. They found that the values of input and output variables obtained by OvMSC, as well as the values of the response variables, belonged to the database available in the ERP system of the company. These results showed that the procedure proposed herein provided feasible and useful solutions to improve the industrial processes under study.