Applying an Electromagnetism-Like Algorithm for Solving the Manufacturing Cell Design Problem

Applying an Electromagnetism-Like Algorithm for Solving the Manufacturing Cell Design Problem

Jose M. Lanza-Gutierrez (Universidad Politecnica de Madrid, Spain), Ricardo Soto (Pontificia Universidad Católica de Valparaíso, Chile), Broderick Crawford (Pontificia Universidad Católica de Valparaíso, Chile), Juan A. Gomez-Pulido (University of Extremadura, Spain), Nicolas Fernandez (Pontificia Universidad Católica de Valparaíso, Chile) and Carlos Castillo (Pontificia Universidad Católica de Valparaíso, Chile)
DOI: 10.4018/978-1-5225-5643-5.ch051

Abstract

Group technology has acquired a great consideration in the last years. This technique allows including the advantages of serial production to any manufacturing industry by dividing a manufacturing plant into a set of machine-part cells. The identification and formation of the cells are known as the Manufacturing Cell Design Problem (MCDP), which is an NP-hard problem. In this paper, the authors propose to solve the problem through a swarm intelligence metaheuristic called ElectroMagnetism-like (EM-like) algorithm, which is inspired by the attraction-repulsion mechanism of particles in the context of the electromagnetic theory. The original EM-like algorithm was designed for solving continuous optimization problems, while the MCDP is usually formulated by assuming a binary approach. Hence, the authors propose an adaptation of this algorithm for addressing the problem. Such adaptation is applied for solving a freely available dataset of the MCDP, obtaining competitive results compared to recent approaches.
Chapter Preview
Top

Background

There are two principal lines of research for papers addressing the MCDP. On the one hand, works considering exact techniques and on the other hand, authors assuming approximate techniques.

Beginning with the first line, there are authors considering different types of exact algorithms. Linear programming approaches were assumed by Purcheck (1975), Oliva-Lopez and Purcheck (1979), and Elbenani and Ferland (2012). Linear quadratic models were considered by Kusiak and Chow (1987) and Boctor (1991). Dynamic programming was assumed by Steudel and Ballakur (1987). Goal programming was assumed by Sankaran and Rodin (1990) and Shafer and Rogers (1991). Constraint programming and boolean satisfiability were considered by Soto et al. (2012). Mixed integer linear programming models were proposed by Krushinsky and Goldengorin (2012) and Fahmy (2015).

Complete Chapter List

Search this Book:
Reset