You are on page 1of 5

Available online at www.sciencedirect.

com

ScienceDirect
Procedia Computer Science 74 (2015) 181 185

International Conference on Graph Theory and Information Security

Identication of Bottleneck Elements in Cellular Manufacturing


Problem
S. Arumugam, J. Saral
National Centre for Advanced Research in Discrete Mathematics,
Kalasalingam University, Anand Nagar, Krishnankoil-626126, Tamil Nadu, India

Abstract
The cell formation problem which arises in cellular manufacturing can be formulated in graph theoretic terms. The input for a
cellular manufacturing problem consists of a set X of m machines and a set Y of p parts and an m p matrix A = (ai j ), where ai j = 1
or 0 according as the part p j is processed on the machine mi . This data can be represented as a bipartite graph G with bipartition
X, Y and mi is joined to p j if ai j = 1. Let G1 , G2 , . . . , Gk be nontrivial connected subgraphs of G such that V(G1 ), V(G2 ), . . . , V(Gk )
forms a partition of V(G). Then = {G1 , G2 , . . . , Gk } is called a k-cell partition of G. Any edge of G with one end in Gi and the
other end in G j with i  j represents an intercellular movement of a part. The set V(Gi ) is called a cell. Any part in a cell which
is to be processed by a machine in another cell is called a bottleneck part and any machine which processes a part in another cell
is called a bottleneck machine. In this paper we use graph theoretic techniques to develop an ecient algorithm for identifying
bottleneck elements with respect to a given solution of the cellular manufacturing problem. If H is the subgraph of G induced by
the set of all exceptional edges with respect to the cell partition , then any vertex of H is a bottleneck element and its degree in H
is the strength of the bottleneck elements. Thus the bottleneck elements of high strength identied by our algorithm are to be either
duplicated if it is a machine or to be subcontracted if it is a part, in order to ensure a smooth ow in the manufacturing process.
c 2015

2015The
TheAuthors.
Authors.
Published
by Elsevier

Published
by Elsevier
B.V.B.V.
This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review
under responsibility of the Organizing Committee of ICGTIS 2015.
Peer-review under responsibility of the Organizing Committee of ICGTIS 2015

Keywords: Cellular manufacturing problem, cell, bottleneck part, bottleneck machine, bipartite graph, depth rst search.
2010 MSC: 94C15, 05C21

1. Introduction
Cellular manufacturing is an application of the principles of group technology in manufacturing. The input for a
cellular manufacturing problem consists of a set X of m machines, a set Y of p parts and an m p matrix A = (ai j ),
where ai j = 1 or 0 according as the part p j is processed on the machine mi . Those parts which require a similar
manufacturing process are grouped into a family, called a part family. Given a part family, a group of machines
is identied for manufacturing the parts of the family and the part family along with the corresponding group of
machines is called a cell. Thus a cell is a small scale, well-dened production unit within a large factory, which
E-mail address: s.arumugam.klu@gmail.com, saral2184@gmail.com

1877-0509 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Organizing Committee of ICGTIS 2015
doi:10.1016/j.procs.2015.12.097

182

S. Arumugam and J. Saral / Procedia Computer Science 74 (2015) 181 185

has the responsibility for producing a family of parts. Cellular manufacturing problem is to design cells in such a
way that some measure of performance is optimized. One of the objectives of the cellular manufacturing problem
is to minimize the number of part movements from one cell to another cell. Given a solution {C1 , C2 , . . . , Ck } for a
cellular manufacturing problem, any part in a cell which is to be processed by a machine in another cell is called a
bottleneck part and any machine which processes a part in another cell is called a bottleneck machine. In this paper
we use graph theoretic techniques to develop an ecient algorithm for identifying bottleneck elements with respect
to a given solution of the cellular manufacturing problem. Any edge of the graph G with both its ends as bottleneck
elements is called an exceptional edge.
By a graph G = (V, E) we mean a nite, undirected graph with neither loops nor multiple edges. The machine-part
incidence matrix A = (ai j ) of a cellular manufacturing problem can be represented as a bipartite graph G = (V, E)
where V = X Y and a machine mi X is joined to part p j Y if ai j = 1. The following example depicts the
representation of the machine-part incidence matrix A = (ai j ) of a cellular manufacturing problem as a bipartite graph
G = (V, E).
Example 1.1.
m1
m2
A = m3
m4
m5
p1
u

p1
1
1
0
0
0

p2
0
0
1
1
1

p2
u

p3
u

u
m1

u
m2

p3
0
1
1
0
0

p4
0
0
0
0
1
p5
u

p4
u

u
m3

p5
0
0
0
1
0

u
m4

p6

0
0
p6
u

u
m5

Fig. 1. The bipartite graph of the matrix A

Many approaches and algorithms have been suggested by various authors for identifying bottleneck elements with
respect to a given solution for a cellular manufacturing problem. Chow and Hawaleshka [4] developed an algorithm
to reduce the number of bottleneck parts. Chan and Milner [3] , King and Nakoranchai [7] , Kern and Wei [6] , Tsai, Chu
and Batra [11] , Won [12] have developed algorithms to reduce the appearance of the bottleneck parts and bottleneck
machines. The standard technique used in the literature to reduce the number of exceptional edges is to duplicate a
bottleneck machine and subcontracting bottleneck parts. Seifoddini [9] presented a cost-based duplication procedure,
which uses the duplication cost and the associated reduction in intercellular material handling cost as basis for decision
making in the duplication process. Ang [1] developed an algorithm for eliminating bottleneck parts that minimizes
total duplication costs for the entire system and its duplication process. Kusiak and Chow [8] proposed subcontracting
bottleneck parts if bottleneck part production cost due to intercellular movements is greater than subcontracting cost.
Shafer, Kern and Wei [10] proposed that a remainder cell may be designed to which the subcontracted parts can be
assigned. Gunasingh and Lashkari [5] employed a cost-based integer programming approach to allow for both machine
duplication and the possibility that some intercellular movements remain.
2. Main Results
Let G be the bipartite graph associated with the machine-part incident matrix of a cellular manufacturing problem.
Let {C1 , C2 , . . . , Ck } be a solution to the problem and let Gi be the subgraph of G induced by Ci . Any edge of G having

S. Arumugam and J. Saral / Procedia Computer Science 74 (2015) 181 185

one end in Gi and other end in G j with i  j is an exceptional edge. Let H denote the subgraph of G induced by the set
of all exceptional edges of G. Clearly any vertex of H is a bottleneck element. The strength of a bottleneck element is
dened to be its degree in the subgraph H. We illustrate these concepts with an example.
Example 2.1. Consider the following cellular manufacturing problem with machine-part incidence matrix

m1
m2
B = m3
m4
m5

p1
1
1
1
0
0

p2
1
1
1
0
0

p3
0
0
1
1
1

p4
0
0
0
1
0

p5
0
0
1
1
1

p6

1
1

The corresponding bipartite graph is given in Fig. 2.


p1
u

p3
u

p2
u

u
m1

u
m2

p5
u

p4
u

u
m3

u
m4

p6
u

u
m5

Fig. 2. Bipartite graph corresponding to the matrix B.

A solution to this problem with two cells is given by C1 = {m1 , m2 , p1 , p2 } and C2 = {m3 , m4 , m5 , p3 , p4 , p5 , p6 }.
The exceptional edges are m3 p1 and m3 p2 . The subgraph H of G induced by the set of exceptional edges is given in
Fig. 3.
m3
s

s
p1

s
p2

Fig. 3. The graph H induced by exceptional edges.

Hence for the given solution m3 is a bottleneck machine with strength two and the parts p1 and p2 are bottleneck
parts with strength one. If we duplicate the machine m3 and place it in the cell C1 , the number of exceptional edges
becomes zero.
We now present an algorithm for identifying bottleneck machines and bottleneck parts for a cellular manufacturing
problem.
Algorithm
Input: Machine part incidence matrix B = (ai j ) and the cells C1 , C2 , . . . , Ck .
Step 1:
Step 2:
Step 3:
Step 4:

S =
If (ai j ) = 1 and mi and p j are in dierent cells, then replace S by S {mi p j }
Repeat Step 2.
Output S

183

184

S. Arumugam and J. Saral / Procedia Computer Science 74 (2015) 181 185

We now construct the subgraph H of G which is induced by the set of all edges in S . Any vertex of H is a
bottleneck element and its degree is the strength of the bottleneck element. If a bottleneck machine has large degree,
then duplicating that machine will reduce the number of exceptional elements. Similarly subcontracting bottleneck
parts with large degree will reduce the number of exceptional elements. We illustrate this algorithm with another
example.
Example 2.2. Consider the following cellular manufacturing problem with machine-part incidence matrix given in
Boe and Cheng [2] .
p1 p2 p3 p4 p5 p6 p7 p8 p9 p10 p11 p12 p13 p14 p15 p16 p17 p18 p19 p20 p21 p22 p23 p24 p25 p26 p27 p28 p29 p30 p31 p32 p33 p34 p35
m1

m2

m3

m4

m5

m6

m7

m8

m9

m10

m11

m12

m13

m14

m15

m16

m17

m18

m19

m20

A solution to the above problem into four cells is given by


C1 = {m11 , m12 , m15 , m16 , m19 , p4 , p6 , p9 , p11 , p21 , p28 , p30 , p32 , p33 },
C2 = {m5 , m6 , m9 , m10 , m20 , p26 , p8 , p14 , p16 , p19 , p22 },
C3 = {m2 , m4 , m13 , m14 , m18 , p2 , p7 , p10 , p12 , p13 , p18 , p24 , p27 , p31 } and
C4 = {m1 , m3 , m7 , m8 , m17 , p1 , p3 , p5 , p15 , p17 , p20 , p23 , p25 , p29 , p34 , p35 }.
The subgraph H induced by the set S of all exceptional edges determined by the above algorithm is given in Fig 4.
p5 p6 p7 p9 p11 p12 p14 p16 p17 p18 p19 p20 p22 p23 p24 p25 p26 p30 p31 p32 p35
r
r
r
r
r
r
r
r
r
r r r
r
r
r
r r r
r r r
D
D
D
D
D
Dr
r
r
r
r
r
r
r
r
r
r
r
m18
m20
m
m
m16
m
m
11
m
9
m3
m5 m6
m1
8
2
7

Fig. 4. The graph H induced by exceptional edges.

The vertices of H are the bottleneck elements. For example m1 is a bottleneck machine whose strength is seven.
Four edges come from the cell C1 , two edges from C3 and one edge from C2 . Hence if the machine m1 is duplicated
and placed in C1 then the number of exceptional elements is reduced by four.
In general decision on duplicating a bottleneck machine or on subcontracting a bottleneck part is a complicated
issue. One has to consider the cost involved in duplicating a bottleneck machine or in subcontracting a bottleneck

S. Arumugam and J. Saral / Procedia Computer Science 74 (2015) 181 185

part and the corresponding reduction in the production cost. For this purpose the entire problem has to be formulated
with further details such as cost of a machine, cost of production of a part, cost involved in subcontracting a part and
increase in prot if a machine is duplicated. A detailed investigation of this problem is our future plan.
3. Conclusion and Scope
In this paper we have developed an algorithm for identifying bottleneck machines and bottleneck parts in a cellular
manufacturing problem along with a given cell partition. The algorithm can be further rened by taking into account
the cost associated with the problem. Results in this direction with practical applications will be reported in subsequent
papers.
References
1. Ang DS. An algorithm for handling exceptional elements in cellular manufacturing systems. Industrial management and data systems
2000;100:251254.
2. Boe WJ, Cheng CH. A close neighbour algorithm for designing cellular manufacturing systems. International Journal of Production Research
1991;29:2097-2116.
3. Chan HM, Milner DA. Direct clustering algorithm for group formation in cellular manufacturing. Journal of manufacturing systems 1982;1:6476.
4. Chow WS, Hawaleshka O. An ecient algorithm for solving machine chaining problem in cellular manufacturing. Computer and Industrial
Engineering 1992;22:95-100.
5. Gunasingh KR, Lashkari RS. Machine grouping problem in cellular manufacturing systems: An integer programming approach. International
Journal of Production Research 1989;27:1465-1473.
6. Kern GM, Wei JC. The cost of elimination exceptional elements in group technology cell formation. International Journal of Production
Research 1991;29: 1535-1547.
7. King JR, Nakornchai V. A machine-component group formation in group technology: A review and extension. International Journal of Production Research 1982;20:117-133.
8. Kusiak A, Chow WS. Ecient solving of the group technology problem. Journal of Manufacturing Systems 1987;6:117-124.
9. Seifoddini H. Duplication process in machine cells formation in group technology. IIE Transactions 1989;21:382-388.
10. Shafer S, Ken G, Wei J. A Mathematical programming approach for dealing with exceptional elements in cellular manufacturing. International
Journal of Production Research 1992;30:1029-1036.
11. Tsai CC, Chu CH, Barta T. Analysis and modeling of a manufacturing cell formation problem with fuzzy integer programming. IIE Transactions 1997;29:533-547.
12. Won Y. Two-phase approach to GT cell formation using ecient p-median formulations. International Journal of production Research
2000;38:1601-1613.

185