Selecting machines and buffers in unreliable assembly/disassembly manufacturing networks Nabil Nahas a,b , Mustapha Nourelfath b,c,n , Michel Gendreau b,d a King Fahd University of Petroleum & Minerals, Systems Engineering Department, Dhahran 31261, Saudi Arabia b Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation (CIRRELT), Canada c Department of Mechanical Engineering, Faculty of Science and Engineering, Laval University, Canada d Department of Mathematics and Industrial Engineering, École polytechnique de Montréal, Canada article info Article history: Received 18 July 2013 Accepted 11 April 2014 Available online 19 April 2014 Keywords: Assembly/disassembly Design Optimization Production rate Genetic algorithms abstract This paper formulates an optimal design model for assembly/disassembly manufacturing networks. The objective is to maximize production rate subject to a total cost constraint. Machines are chosen from a list of products available on the market, and sizes of the buffers are chosen within a predetermined range. Each machine type is characterized by its total cost of ownership, failure rate, repair rate and processing time. The buffers are also characterized by their total cost of ownership coefficients associated with the buffer size. To estimate assembly/disassembly network performance, a decomposition-type approximation is used. The optimal design model is formulated as a combinatorial optimization one in which the decision variables are buffers and types of machines. A genetic algorithm is proposed as an optimization technique. Numerical examples are used to highlight the benefit of selecting simultaneously the buffers and the machines. & 2014 Elsevier B.V. All rights reserved. 1. Introduction 1.1. Problem description A manufacturing system can be defined as a set of machines, storage buffers, conveyers, transportation components, computers, and other elements that are used together for manufacturing (Gershwin, 1994). In manufacturing systems, assembly operations consist of bringing two or more parts together to create a single product, while disassembly operations consist of separating a product into two or more parts. A large number of manufacturing systems can be modeled as high-volume, assembly/disassembly (A/D) networks of unreliable machines separated by finite buffers. Such A/D networks are widely encountered in industry (Burman, 1995). Assembly operations are frequently encountered for exam- ple in automotive, electronics, window and door industries. Even if assembly seems to be more important in manufacturing, disas- sembly operations also occur widely, for example in cloth cutting, sheet metal cutting, recycling, and waste handling (Kouikoglou, 2002). Furthermore, assembly and disassembly operations can be used to model transfer mechanisms in which parts are loaded on pallets where they undergo a set of operations and, upon comple- tion of these operations, the parts are unloaded and the pallets are liberated (Di Mascolo et al., 1991). Fig. 1 illustrates an acyclic A/D network. Squares represent machines or stations while circles represent buffers. A square can also represent a sequence of machines without buffers, since in the following we treat a sequence of machines without buffers as a single machine. At each station, machines are denoted by M 1 , M 2 , etc. An intermediate buffer connecting one upstream machine M k and one downstream machine M i is denoted by B k, i . As shown in Fig. 1, material flows in the direction of the arrows. Machines M 3 and M 4 are fed by infinite sources of raw parts and machines M 1 and M 2 release items into infinite output buffers. Machine M 3 disassembles a part into two sub-products that are sent to machines M 2 and M 1 for further processing, whereas M 2 assembles two parts into a composite product. It takes one part in each of M 2 upstream buffers B 4,2 and B 3,2 to perform an assembly operation. At the disassembly machine M 3 , one part is needed from its upstream buffer to perform the disassembly operation. It is advantageous to find ways to optimize machine selection and buffer space allocation to make factories more efficient and more profitable. Tools for rapid design of manufacturing systems are especially important for products with short life cycles. Usually, designers of such systems want to optimize the produc- tion rate of the system. However, material flow may be disrupted by machine failures. Both the selection of more efficient machine Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/ijpe Int. J. Production Economics http://dx.doi.org/10.1016/j.ijpe.2014.04.011 0925-5273/& 2014 Elsevier B.V. All rights reserved. n Corresponding author at: Department of Mechanical Engineering, Faculty of Science and Engineering, Laval University, Quebec (QC.), G1V 0A6, Canada. Tel.: þ1 418 6562131x12355; fax: þ1 418 6567415. E-mail addresses: Nabil.Nahas.1@ulaval.ca (N. Nahas), Mustapha.Nourelfath@cirrelt.ca (M. Nourelfath), Michel.Gendreau@cirrelt.ca (M. Gendreau). Int. J. Production Economics 154 (2014) 113–126