Computers ind. Engng Vol. 19, Nos 1-4, pp. 294-298, 1990 0360-8352/90 $3.00 + 0.00
Printed in Great Britain. All rights reserved Copyright © 1990 Pergamon Press pie
MODELING THE MACHINE LAYOUT PROBLEM
Sunderesh Heragu
Department of Management and Marketing
School of Business and Economics
State University of New York
Plattsburgh, NY 12901
ABSTRACT
using a modified penalty algorithm. The
In this paper, a mathematical model for the algorithm produced solutions of high quality
machine layout problem is presented. The and requires low computation time. In this
objective function of the model minimizes the paper, a new model for the machine layout
total cost involved in transporting material problem is presented. The model is an
between each pair of machines. The extension of the one presented in Heragu
constraints ensure that: (i) the pickup and (1988). To solve the model presented, a
drop-off points of a machine fall within the simulated annealing algorithm is also
boundaries of the machine~ (ii) no two presented.
machines in the layout overlap; and (iii) the
machines are located inside the building. A In the next section, some of the models
simulated annealing algorithm for solving the developed for the facillty/machine layout
model is also presented. The solution problem are listed. In section 3, models for
provided by the algorithm indicates the MHS the single-row and multl-row machine layout
paths, the relative positioning of each problem are presented. A simulated annealing
machine, the pick-up and drop-off points for algorithm for solving the layout problem is
each machine, and clearance between machines, also presented. Concluding remarks and some
llmltations of the simulated anneallng
algorithm are provided in section 4.
INTRODUCTION
The objective of developing machine layout in LITERATURE REVIEW
an automated manufacturing system is to
minimize the total time required by the Traditionally, the layout problem has been
material handling systems to transport modeled as a QAP. Given that:
material between the machines. Two patterns n total number of machines or locations
of machine layout frequently occur in fik number of trips to be made between
practice (Heragu and Kusiak, 1990): machines i and k
- slngle-row machine layout Cjl cost per trip between locations j and 1
- multi-row machine layout, xij is I if plant i is at location j;
0 otherwise
The single-row and multi-row layout
problems have also been referred to as one- the QAP can be stated as (Koopmans and
dimensional and two-dimensional space Beckman, 1957):
allocation problems (Simmons, 1969). The
single-row (multi-row) layout problem n n
involves problems in which the machines are Min ~ ~ flkCjlXijXkl (I)
arranged linearly along one row (two or more i-I j=l
rows). Heragu and Kusiak (1990) identify a
number of applications of the single-row and n
multi-row layout problems, s.t. ~ xij - l, i=l,...,n (2)
J=l
In this paper, the development of
machine layout in a manufacturing system is n
discussed. Traditionally, the layout problem ~ xij - i, j-l,...,n (3)
has been modeled as a quadratic assignment i-I
problem (QAP). It is well-known that the QAP
is NP-complete and cannot be used to solve x~j - 0 or l, i,j-I .... ,n (4)
large scale layout problems, for example,
problems in which there are fifteen or more A number of other models have also been
machines. Heragu (1988) has presented a new developed for the layout problem. For
model with absolute values in the objective example, Bazaraa (1975) presented the
function and constraints and solved the model Quadratic Set Covering model~ Love and Wont
294