This work proposes a new methodology to optimize flexible manufacturing cells applying Simulated Annealing and Contribution Margin concepts. This methodology helps to determine the number of extra products that could be sold and manufactured while taking advantage of this idleness. The price formation to sell the extra products was based on contribution margin concepts. The main concept is to use cell idleness to increase profits while the costs are already paid. Equipment and workers have their fixed costs paid regardless of whether they are active or not. When they are not active, all of their costs are considered to be losses. What motivated the present work was the fact that, in general, the idleness detected in manufacturing processes does not receive adequate attention in the sense of taking advantage of it. To achieve this purpose, a program was developed to implement the methodology. This program generates user interfaces that allow the prediction of daily scheduling parts production. The first interface identifies only bottlenecks and idleness. The second interface, based on the Simulated Annealing (SA) metaheuristic technique, was designed to perform the management of idleness. A database was organized to present current and future daily scheduling parts production. Based on the volume of sales and the lead time of sold products, sellers are always able to update this database. Thus, the program can provide through its user interface, at any time, all of the information about production data for a specific schedule day that comes from the database. The user mentioned here is expected to be an industry salesman. Thus, it will always be possible to have previous knowledge about the bottlenecks and idleness that exist for any scheduled day in the future. This knowledge allows for planning the process of selling extra products made from the existing idleness. Two aspects were considered with respect to selling extra products that are produced via taking advantage of the idleness: the delivery time and the contribution margin of each product. To validate the program, a simulation was performed to determine which products should be produced during a scheduled day to ensure the delivery times or to ensure the maximum contribution margin, while considering the delivery time that is available. As a final consideration, the software can be considered to be functional and responsive to the possibility of profit maximization from the use of cell idleness.
Resumo. Algoritmos de otimização com dois níveis usam, em geral, uma combinação de dois modelos: um definido no espaço de busca original (fine model) e outro em um subespaço aproximado (coarse model). Em alguns casos, o coarse model precisa ser refinado a cada iteração elevando o custo computacional. Este trabalho apresenta um Algoritmo Genético de dois níveis que utiliza um coarse model global definido com base em um modelo de rede neural artificial, o qualé criado a partir da análise de componentes principais de um conjunto preliminar de soluções obtidas no espaço de busca original. Para ilustrar uma aplicação da abordagem proposta são apresentados os resultados promissores para um problema exemplo.Palavras-chave. Algoritmo Genético, Otimização, Análise de Componentes Principais, Métodos dois-níveis, Rede Neural Artificial. IntroduçãoA aplicação de algoritmos de otimização metaheurísticos, como o Algoritmo Genético (AG),à problemas das engenhariasé muito difundida [1][2][3]. De modo geral, as abordagens baseiam-se na construção de um modelo substituto (surrogate model) construído a partir de um conjunto limitado de simulações computacionais, por exemplo, do método dos elementos finitos. A fim de encontrar uma soluçãoótima, ou próxima doótimo, esses algoritmos normalmente exigem muitas avaliações do modelo, o que demanda um custo computacional razoavelmente alto.Muitos algoritmos de dois níveis tem sido proposto na tentativa de acelerar a resolução desses problemas de otimização [4,5]. Esses algoritmos utilizam uma combinação de dois modelos substitutos no procedimento de otimização: um definido no espaço de busca original (fine model) e outro em um subespaço (coarse model). O coarse modelé criado com base em algumas avaliações do fine model e a sua precisão influencia fortemente a convergência. Em muitos casos, porém, os algoritmos de dois níveis possuem características locais, o que significa que o coarse model deve ser construído a cada iteração o que representa uma importante desvantagem dessas abordagens.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.