TECHNICAL NOTE

INTEGRATED APPROACH FOR CELLULAR

MANUFACTURING – A CASE STUDY

B. S. Nagendra Parashar and Ramprasad Kamath

Mechanical Engineering Group, Birla Institute of Technology and Science

Pilani, Dubai Campus, Dubai, UAE, bsnagendraa@yahoo.co.in

H.V. Somasundar

Sri Krishna Institute of Technology

Bangalore, India, hvs-sundar@yahoo.com

(Received: August 12, 2002 – Accepted in Revised Form: December 8, 2003)

Abstract To cope with fast changing customer requirements, industrial demands and to meet
stringent specifications of customers Cellular Manufacturing Systems has become an effective tool in
hands of manufacturers. Most of the published literature on cell formation earlier considers only the
data available in the route sheets and ignored subproblems associated with cell formation. There is a
need to develop an integrated approach for Cellular Manufacturing, which considers all the aspects.
Present paper focuses on developing an integrated approach for Cellular Manufacturing. Proposed
method considers workload, volume of production and processing times for machine cell and part
family formation, which was ignored by past researchers. Proposed method has been tested on one of
the largest public sectors in Asia and the benefits that can be obtained by implementation are
reported.

Key Words Cellular Manufacturing, Machine Cells, Part Family, Exceptional Component,
Workload Index, Cellular Layout

‫ﭼﻜـﻴﺪﻩ ﺑـﻪ ﻣـﻨﻈﻮﺭ ﭘﺎﺳـﺨﮕﻮﻳﻲ ﺑـﻪ ﺗﻐﻴﻴﺮﺍﺕ ﺳﺮﻳﻊ ﺩﺭ ﺗﻘﺎﺿﺎﻱ ﻣﺸﺘﺮﻱ ﻭ ﺩﺭﺧﻮﺍﺳﺖ ﺻﻨﻌﺖ ﻭ ﺑﺮﺍﻱ ﺑﺮ ﺁﻭﺭﺩﻥ‬
‫ ﺑﻴﺸﺘﺮ‬.‫ ﺗﻮﻟﻴﺪ ﻛﻨﻨﺪﮔﺎﻥ ﺩﺭﻳﺎﻓﺘﻪ ﺍﻧﺪ ﻛﻪ ﺳﻴﺴﺘﻤﻬﺎﻱ ﺗﻮﻟﻴﺪ ﺣﺠﺮﻩ ﻭﺍﺭ ﺍﺑﺰﺍﺭ ﻣﻮﺛﺮﻱ ﻣﻲﺑﺎﺷﺪ‬،‫ﻣﺸﺨﺼﺎﺕ ﺩﻗﻴﻖ ﻣﻮﺭﺩ ﻧﻴﺎﺯ‬
‫ﻣﻘـﺎﻻﺕ ﭼـﺎﭖ ﺷﺪﻩ ﺩﺭ ﺑﺎﺭﻩ ﺗﺸﻜﻴﻞ ﺣﺠﺮﻩ ﻫﺎ ﺗﻨﻬﺎ ﺑﻪ ﺩﺍﺩﻩ ﻫﺎﻱ ﻣﻮﺟﻮﺩ ﺩﺭ ﺑﺮﮔﻬﺎﻱ ﻣﺴﻴﺮ ﻛﺎﺭ ﻣﻲﭘﺮﺩﺍﺯﻧﺪ ﻭ ﻣﺴﺎﺋﻞ‬
‫ ﺍﺯ ﺍﻳﻦ ﺭﻭ ﻳﻚ ﺩﻳﺪﮔﺎﻩ ﻳﻜﭙﺎﺭﭼﻪ ﻭ ﺟﺎﻣﻊ ﺑﺮﺍﻱ ﺗﻮﻟﻴﺪ ﺣﺠﺮﻩ ﻭﺍﺭ ﻛﻪ‬.‫ﺭﻳﺰ ﭼﮕﻮﻧﮕﻲ ﺗﺸﻜﻴﻞ ﺣﺠﺮﻩ ﻫﺎ ﺭﺍ ﺩﺭ ﺑﺮ ﻧﺪﺍﺭﻧﺪ‬
،‫ ﻳﻌﻨﻲ ﺑﺎﺭ ﻛﺎﺭﻱ‬.‫ ﺍﻳﻦ ﻣﻘﺎﻟﻪ ﺗﻼﺷﻲ ﺩﺭ ﺍﻳﻦ ﺭﺍﺳﺘﺎ ﻣﻲﺑﺎﺷﺪ‬.‫ ﺳﺎﺯﻧﺪﻩ ﺧﻮﺍﻫﺪ ﺑﻮﺩ‬،‫ﺩﺭ ﺑـﺮ ﮔـﻴﺮﻧﺪﻩ ﻫﻤـﻪ ﺟﻮﺍﻧـﺐ ﺑﺎﺷﺪ‬
‫ ﺍﻳﻦ ﺭﻭﺵ ﭘﻴﺸﻨﻬﺎﺩﻱ‬.‫ ﺯﻣﺎﻥ ﻓﺮﺍﻭﺭﺵ ﻣﺎﺷﻴﻦ ﺩﺭ ﻫﺮ ﺣﺠﺮﻩ ﻭ ﺗﺸﻜﻴﻞ ﺧﺎﻧﻮﺍﺩﻩ ﻗﻄﻌﺎﺕ ﺭﺍ ﻣﻨﻈﻮﺭ ﻣﻲﺩﺍﺭﺩ‬،‫ﺣﺠﻢ ﺗﻮﻟﻴﺪ‬
‫ﺭﻭﻱ ﻳﻜـﻲ ﺍﺯ ﺑﺰﺭﮔﺘﺮﻳـﻦ ﺑﺨﺸـﻬﺎﻱ ﻋﻤﻮﻣـﻲ ﺁﺳـﻴﺎ ﺁﺯﻣـﻮﺩﻩ ﺷﺪﻩ ﻭ ﻧﺘﺎﻳﺞ ﺑﺪﺳﺖ ﺁﻣﺪﻩ ﺍﺯ ﻛﺎﺭﺑﺮﺩ ﺍﻳﻦ ﺭﻭﺵ ﺟﺪﻳﺪ‬
.‫ﮔﺰﺍﺭﺵ ﺷﺪﻩ ﺍﺳﺖ‬

1. INTRODUCTION Systems, and Agile Manufacturing. It is a basic

requirement for adopting new manufacturing
Cellular Manufacturing has been gaining importance philosophies like Toyota Production Systems, Lean
for both practitioners and researchers in the area of Manufacturing, Nagare Production Systems etc.
manufacturing during the last decade. It is one of Cellular Manufacturing probably had a greater
the strategies the organization should adopt before impact on increasing manufacturing productivity
implementing new technologies like Computer than any other manufacturing concept. This concept
Integrated Manufacturing, Flexible Manufacturing has helped many organizations to compete more

IJE Transactions A: Basics Vol. 17, No. 1, February 2004 - 73

aggressively at global level by improving The similarity coefficient method lends itself more
manufacturing flexibility, quality and at the same easily to computer applications.
time reducing lead-time and manufacturing costs There is a need to develop a cell formation
simultaneously. technique, which can handle multiple objectives
The basic idea in Cellular Manufacturing is to pertinent to Cellular Manufacturing such as
group parts that have similar processing needs into minimization of work-in-process inventory,
part families, and machines that meet these needs material handling, setup time, maximization of
into machine cells. Each machine cell is dedicated machine utilization etc. Most of the published
to the manufacture of a part family. The advantages literature addresses the problem of cell formation
of Cellular Manufacturing over traditional alone and ignores processing of exceptional
manufacturing are many folds. To mention a components (components required to be processed
few of them reduced setup time, work-in-process in more than one cell), arranging cells and
inventory, lead-time, simplified flow of materials, machines within the cells, production control
improved huma n re l a t i ons e t c . Cellu lar activities etc.
Manufacturing combines the efficiency of a flow The paper presented here is an integrated
shop layout with the flexibility of a job shop approach for Cellular Manufacturing. Present
layout. paper deals with new heuristics along with
The first problem that must be addressed when similarity coefficient for machine cell and
considering a Cellular Manufacturing is that of cell part family formation. Processing of exceptional
formation. The cell formation method is of components and Cellular layout are also dealt with.
paramount importance when implementing a Proposed method is tested on 20-machine 144-
Cellular Manufacturing because the success of component problem in one of the largest public
Cellular Manufacturing depends greatly on the sectors in India.
initial grouping of machines and parts. Ever since
the concept of Cellular Manufacturing is
introduced attempts were made by different
researchers and practitioners to develop algorithms 2. PROBLEM STATEMENT
for the efficient cell formation. Few of them include
the application of classification and coding systems, The problem under consideration is taken from one
flow analysis, similarity coefficient methods, of the largest public sectors in India, which is
mathematical models, graph theory models, genetic regarded as one of Asia’s leading corporate giants.
algorithms, neural networks, fuzzy sets theory etc. The company faced the problems of having
All the cell formation techniques mentioned increased material handling, larger setup times (as
above have their own limitations. Mathematical high as 15-20 times processing time), higher work-
models Srinivasan [1], Crama and Ostan [2], in-process inventory, longer lead-time, queue time,
Jayakrishanan [3] et al. are all characterized by difficulty in meeting delivery schedules etc. To
week continuous relaxations, and hence, are not overcome these problems company resorted to
well suited for the solution to optimality of large- apply Cellular Manufacturing.
scale problems. For large-scale problems their The company is undertaking the manufacture of
usefulness is purely illustrative. Similarity coefficient different projects. Since the introduction of Cellular
method is among the most frequently cited Manufacturing involves substantial investment,
approach to cell formation problems. Several risk and time, it was decided to apply Cellular
measures of similarity among machines and parts Manufacturing for one of the project WA-200.
have been proposed in the last three decades. Based on the results obtained it was decided to
Applications of similarity coefficient to cell extend the application of Cellular Manufacturing to
formation include the contributions of McAuley other projects. WA-200 is a wheel loader having
[4], Waghodekar and Sahu [5], Nagendra Parashar boom and bucket at the front, which is used to
and Somasundar [6] etc. Similarity coefficient carry loads from one place to another. WA-200
methods are more flexible in incorporating consists of 144 components to be processed on 20
manufacturing data into the machine cell formation. machines. Many of these components are precision

74 - Vol. 17, No. 1, February 2004 IJE Transactions A: Basics

 TABLE 1. Components and its Process Sequence for WA-200 Project.

Component Component
Process Sequence Qty./Equipment
No. Name
1 Boss 8-13-2-5-18 1
2 Plug 8-13-2-15-18 1
● ● ● ●
● ● ● ●
● ● ● ●
143 Wheel disc 3-5-18 4
144 Zipped block 17-5-18 8

type having close dimensional tolerances and good processes a particular component otherwise go to
surface finish. The number of manufacturing Step 2.
operations carried out varied from 1 to 8. Process
sequence for the components is shown in Table 1. Step 2 Compute similarity coefficient between
all machines using the formula [6]:

3. METHODOLOGY
SC ij = 1 if i = j
nj

The problem under consideration is having large ∑ (a ip × a jp )

number of components for processing on different SC ij =
p =1
if i ≠ j (1)
nj
machine tools (refer to Table 1). For the reasons
mentioned earlier straightforward application of ∑ (a )
p =1
ip
any of the cell formation methods discussed
earlier might have not resulted in efficient
machine cell. In this paper we have combined aip = 1 if component is processed by the machine,
similarity coefficient with simple heuristics to ‘0’ otherwise.
obtain an optimal solution. Proposed method is an
integrated approach for cell formation, processing j = 1,2,..., nj
exceptional components and arrangement for
Cellular layout. Present work is an extension of i = 1,2,..., nm
work done by Nagendra Parashar and Somasundar
[6]. Proposed method considers workload of where nj = number of jobs and nm = number of
machines, utilization and cost parameters, which machines.
was ignored by past researchers. Proposed method The similarity coefficient varies from 0 and 1.
is explained below. ‘0’ similarity coefficient implies machines within
the cell are purely dissimilar and they do not have
Step 1 Represent the data in the form of machine- a single common machining operation. Similarity
component incidence matrix: coefficient ‘1’ implies machines have complete
Represent the data given in Table 1 in the form common operations.
of machine-component incidence matrix with
machines in row and components in column Step 3 Categorize machines into ideal, critical
position (not shown). Enter ‘1’ if the machine ‘0’ and non-critical machines.

IJE Transactions A: Basics Vol. 17, No. 1, February 2004 - 75

 TABLE 2. Number of Components and Machines for WA-200 Project.

Machine
Name of the Machine No. of Machines No. of Components
No.
1 IC turret lathe 1 1
2 H 22 lathe 3 33
3 HMT L 45 lathe 1 1
4 Bombay lathe 3 1
5 Radial drilling 2 85
6 Bench drill 4 3
7 Circular saw 1 16
8 Horizontal bandsaw 1 11
9 Milling FN 2H 2 3
10 Milling FN 3H 1 1
11 Vertical milling 1 30
12 Pedestal grinder 2 2
13 Centering machine 1 2
14 Facing & Centering 1 10
15 AF7 Boring 2 27
16 WMF Boring 1 5
17 Layout 2 64
18 Bench 2 151
19 A211 Boring 1 1
20 Thread chasing m/c 1 1

TABLE 3. Categorization of Machines machine, vertical milling machine etc.

Categorization Machines
Ideal 2,3,7,8,14,15 Ideal Machine Machines are less in number,
and they process a few components viz. H-22
Non-critical 1,4,6,9,10,12,13,16,19,20
lathe, AF7 boring machine, etc.
Critical 5,11,17,18
Non-Critical Machine Number of machines
available is more and they process very few
components viz. IC turret lathe, Bench drill, etc.
Identify number of components associated with It is clear that the possibility of the formation of
different machines and number of machines a cell with critical machines is abysmally low
available. This is shown in Table 2. because of the very fact that they are associated
Now, categorize machines as critical, non- with a large number of components. Since, Ideal
critical and ideal machines as per the following machines process less number of components and
guidelines from the data available in Table 2. their availability is good they lend themselves to
Critical Machines Only a few machines of this efficient cell formation.
type are available and are associated with processing Histogram for the categorization of machines is
of large number of components viz. radial drilling shown in Figure 1.

76 - Vol. 17, No. 1, February 2004 IJE Transactions A: Basics

 80
Number of Components

70
60
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Machine Number

Figure 1. Histogram for categorization of machines.

2,7,14 15 8 3 of components per machine) and then take the

histogram. The categorization made from the
(a)
histogram is shown in Table 3.

2,7,14,1,4 15,10 8,20,12,13 3,6,9,16,19 Step 4 Obtain initial solution (Machine Cells):
Group the machines with highest similarity
(b) coefficient from the ideal machines group.
Machines 7 and 2 are having similarity coefficient
2,7,14,1,4 15,10,11 8,20,12,13 3,5,6,9,16,19
of 0.93, machines 14 and 2 are having similarity
coefficient of 1.0 and machines 14 and 7 are
having a similarity coefficient of 1.0 (similarity
(c)
coefficient table not shown). Hence, group these
machines in one cell. Machines 15, 8 and 3 are
Figure 2. (a) Core Cell, (b) Revised Cell, (c) Basic Solution.
having zero similarity coefficient with each other
and hence, they are put in different cells. Call this
solution as Core cell. Core cell is shown in Figure
2 (a).
Now, repeat the same procedure by adding non-
The classification is to an extent left to intuition, critical machines from Table 3 to the core cell. Call
though a histogram of the number of machines this solution as revised solution. This is shown in
available and the number of components that are Figure 2 (b).
processed by it greatly aids in this classification. Now, once again revise the solution by adding
To plot this histogram we convert the number of critical machines from Table 3 to the revised
components processed by the machine as if only solution shown in Figure 2 (b). Call this as basic
one of its kinds were available (similar to number solution. This is shown in Figure 2 (c).

IJE Transactions A: Basics Vol. 17, No. 1, February 2004 - 77

TABLE 4. Part Family for WA-200 Project. 144 (number of components). Analyzing such
a big matrix for efficient part family formation
Cell number 1 2 3 4
would be an impossible task. For part family
No. of m/c’s 5 3 4 6 formation uses the following procedure.
No. of comp. 35 29 35 45 Assign points to each component using the
formula

TABLE 5. Relative Workload Index for Different µ i = x i × p j (x i ) (2)

Machines.
where µi = frequency of operations in the ith part
Machine RWLI Machine RWLI family
No. No.
1 0.012 10 0.087 xi = number of operations in ith cell
pj(xi) = probability of operations in the jth job
2 0.93 11 1.00 nj = number of operations in a job
3 0.016 12 0.05
4 0.007 13 0.02 Assign components to that part family where it
5 0.512 14 0.171 has scored highest point.
After applying Equation 2 the number machines
6 0.037 15 0.515 and components associated with each cell is shown
7 0.42 16 0.047 in Table 4.
8 0.05 19 0.016
9 0.002 20 0.016 Step 6 Compute Workload Index and duplicate
additional machines based on workload index and
cost parameters.
Workload index for machine I (WLI) is
Note that in the initial solution (Figure 2 (c)) we calculated using the formula:
have considered only one machine of each type,
though more than one machine is available for a Work load index = WL1 =
few types of machines like Radial drilling  Number of machine hours for 
machine, Milling machine FN 2Hetc (refer Table   (3)
3). These additional machines are assigned to the  which machine is engaged 
initial solution obtained and additional machines Number of hours available
are duplicated wherever required. While assigning
additional machines to the cell(s), we treated the Company works 2 shifts a day (16 hrs.) and 300
problem as if the machine is being duplicated. working days a year.
This is to justify the assigning of the machine(s) WLI greater than ‘1’ implies that the machine
into different cell(s). If the assigning of additional is associated with larger processing times or
machine(s) is not justified, then we can make use number of machines available is insufficient. If
of the same machine(s) available for other projects WL I is less than ‘1’, it implies machine is
the company is undertaking. This is shown in step underutilized. If WLI = 1 then machine is utilized
6. for 100% of its efficiency. The limitation of
workload index is that the index can vary from
anything slightly greater than ‘0’ to very large
Step 5 Formation of Part family: We can
number (theoretically speaking infinitely large).
m a k e u s e o f E q u a t i o n 1 w i t h s l i gh t By dint of this, it is not possible to ascertain as to
modification for part family (jobs having which machine is to be considered for
similar manufacturing attributes and processed duplication. To overcome this problem Relative
in common cell) formation as well. It results in workload index is calculated. Relative workload
similarity coefficient matrix of the size 144 × index of Ith machine (RWLI) is calculated using

78 - Vol. 17, No. 1, February 2004 IJE Transactions A: Basics

 G2 G1 TABLE 7. Table for Assigning Points for Machines.

Cost of the machine (Rs.) Points

<= 1,00,000 7
1,00,000 < cost <= 2,00,000 5
2,00,000 < cost <= 3,00,000 4
Frequency

2
3,00,000 < cost <= 4,00,000 3
1
≥ 4,00,000 1

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Relative workload index To decide about the machines to be duplicated
use the following procedure
Figure 3. Frequency vs relative workload index. (a) Plot a graph of relative workload index (x-
axis) vs. frequency of occurrence (y-axis) as
shown in Figure 3.
(b) Identify gaps (G1,G2,…,Gn) that segregate
TABLE 6. Cost details of Different Bottleneck Machines. various block of workload index. If gap G1 is
chosen, it implies machines having workload
Cost index greater than 0.8 (refer Figure 3) shall
M/C Name of the
(Rs. Life(yrs.) only be duplicated. One can consider the gap
No. m/c
Lac) G2 also. But, this requires more investment
1 IC turret lathe 3.85 10 than earlier case as more number of machines
2 H22 lathe 2.5 10 is considered for duplication (discussed in the
5 Radial drilling 4.15 10 latter part of the paper).
7 Circular saw 0.5 5 (c) List all the machines considered for
9 Milling FN2H 2.6 7 duplication.
Vertical (d) Now, calculate cost workload index for
11 1.8 10
milling machines, which are considered for duplication
Centering using the equation
13 0.3 5
machine
Facing and CWL I = RWL I × Po int s (5)
14 1.5 4
Centering
15 AF7 Boring 2.35 10
The ‘points’ in Equation 5 depends upon the
cost of the machine. Table 6 shows costs of
bottleneck machines (machines associated with
many exceptional components) and their expected
life.
the relation ‘Points’ in Equation 5 are determined by cost of
the machine and slab points. It is shown in Table
Re lative Work load index = RWL 1 = 7. Points given in Table 7 may vary depending
upon range of costs of the machine(s) involved,
Workload index of machine I (4) budgetary constraint and number of machines
Maximum workload index short-listed for duplication. From the Table 7 it is
clear that greater the cost of the machine, lesser
Relative workload index lies between 0 and 1. will be its chances for duplication.
Relative workload index for different machines is (e) Arrange all machines in ascending order of
shown in Table 5. cost workload index.

IJE Transactions A: Basics Vol. 17, No. 1, February 2004 - 79

(f) Chose Ith machine for duplication if it satisfy obsolescence of existing machineries, equipments
the criterion and is a time consuming process. Hence these
options were also not considered. We opted for
eliminating the exceptional components by
k 
CWL I ≥ θ × ∑ CWL I  (6) duplicating the bottleneck machines considering
 I =1  workload, cost, its associativity with exceptional
components and justification for investment.
From the basic solution obtained above (Figure
where k = number of machines considered for 2 (c)) 45 components were found to be exceptional
duplication. θ is a constant which ranges from 0.2 out of 144 components. Batch quantity for these
– 0.4 and depends on the budgetary constraints. exceptional components varied between 200-800.
Lower the θ, the larger the number of machines Among 45 exceptional components 8 components
considered for duplication. It should be noted here required processing in 3 cells and 37 components
that all the machines, which qualify this criteria, required processing in 2 cells. For duplication
need not be duplicated. Final duplication of the and/or assignment of additional machines into
machines is decided by the factors such as cells our major emphasis was to consider
budgetary constraint, number of exceptional workload, its associativity with exceptional
components processed by the machine, machine components, limitations of budgetary constraints
utilization and workload after duplication. and justification for investment. It should be noted
This step is emphasized in more detail in that facilities 17, 18 (Layout and Bench) are
processing the exceptional components. required by all the cells and hence, it was decided
to keep these facilities as common facility (Refer
Step 7 Process exceptional components Figure 5).
Nevertheless the exceptional component, which
required to be processed in more than one cell Considering Workload Select the gap G1 in
often, makes the task of engineers difficult and the Figure 3 for duplication. There are only 2
solution to formation of cells become nontrivial. machines (2 and 11) in this category, which has
To realize the complete benefits of Cellular very high relative workload index (0.93 and 1
Manufacturing, it is imperative to have minimum respectively). Considering gap G1, we calculate
or if possible zero exceptional components. cost workload index for machines 2 and 11 using
Because, presence of exceptional component Equation 5.
increases material handling, setup time, work-in-
process inventory and make the job of scheduling CWLI2 = 0.93 × 4 = 3.72 and CWLI11 = 1 × 5 = 5 .
more difficult and ultimately reduces productivity.
But in real life situations we can find completely Choosing θ = 0.3 in Equation 6 we find both
independent cells only in illustrative examples. machines qualify for duplication.
Many algorithms proposed in the available Repeating the same procedure in gap G2
literature for processing exceptional components (machines 5, 7, and 15), machines that qualify for
assume that cell formation is already done. duplication are 7 and 15.
Exceptional components can be eliminated by Hence, machines, qualified on the basis of
duplicating the bottleneck machines (machines workload index for duplication, are 2, 7, 11 and 15.
associated with many exceptional components), Final decision on duplication of these machines
subcontracting the part, changing manufacturing will be made considering the associativity of these
process, redesigning the part, forming remainder machines with exceptional components and
cell (cell dedicated to the manufacture of justification for duplication.
exceptional components only). Owing to some
extraneous reasons, the organization has not Associativity with Exceptional Components
considered the option of sub-contracting the part. Diagrammatic representation of the exceptional
Changing manufacturing process or redesigning element is shown in Figure 4. It represents the
the part requires substantial investment, involves machines involved and also the number of

80 - Vol. 17, No. 1, February 2004 IJE Transactions A: Basics

 Machine 1
1

Machine 2 11 6 Machine 11

3
Machine 15 5
6 2
Machine 11 Machine 15

3
Machine 14 (a)
Fi 4( )
1 Machine 5

13
Machine 8 2
1 Machine 11

(b)
Machine 2
6

1 Machine 2

1
Machine 11 5
1 Machine 15

2
Machine 14 (c)
Fi 4( )
Machine 13 2

9
2 Machine 5

8
Machine 8

(d)

Figure 4. Diagrammatic representation of exceptional component.

exceptional components that flow between them. consideration contributes exceptional components.
In Figure 4 the number in the circle represents The machine numbers are written at the start or
the machine under consideration. The arrows end of the arrow and the frequencies are written
pointing towards it are indicators of the machines along the arrow. Machines identified as bottleneck
that contribute exceptional components to the are shown in Table 6. It should be noted that all the
machine under consideration. The outward arrows machines, which qualified for duplication based on
show the machines to which the machine under workload index criteria are all listed in Table 6.

IJE Transactions A: Basics Vol. 17, No. 1, February 2004 - 81

Justification for Duplication Duplication of A machine is not considered for duplication if
machine(s) is an important decision that an NPV<0, or BCR<1 or NBCR<1. In any case, if the
organization has to take. The reason is, it calls for value of NPV, BCR and NBCR criteria is satisfied
investment and in the long run, significant savings for more than one machine while eliminating
in material handling costs resulting from decreased exceptional component than the machine(s) with
intercellular movements need to be realized. higher value of NPV, BCR, NBCR and which
Various factors that need to be considered while satisfy the budgetary constraints will be chosen.
duplicating the machine(s) include cost of the We calculate the cash flow (CF) for our
machine, its associability with exceptional calculations using the equation
components, and reduction in intercellular moves
after duplication etc. In the present work cash flow CF = (Number of int ercellular movements)
(CF) is being considered for justifying the × (No. of components)
investment made. Among the different alternatives
available for calculating cash flows Net Present × (Total cos t of moveing the part )
Value (NPV), Benefit Cost Ratio (BCR) and Net (10)
Benefit Cost Ratio (NBCR) has been made use of.
The data required for calculating the cost of
Brief outline of the criteria considered is discussed
moving the part (not shown) was obtained from
here.
Industrial Engineering Department.
The Net Present Value rule states that ‘an
alternative should be adopted only if the present
value of the cash flow generated in the future Alternatives for the Elimination of
exceeds its cost’. NPV is the net profit that accrues Exceptional Components Looking into the
to the firm from adopting the investment. Figure 4a we can see that machines 5 and 11 are
Assuming a constant cash flow Net Present Value bottleneck machines. Considering this we have the
is given by option
1. Duplicate machine 5 in cell 2
NPV = CF × (CRF, CC, n ) − I (7) 2. Duplicate machine 11 in cell 4

where: Considering the 1st option, i.e. duplicating

machine 5 in cell 2, we calculate NPV, BCR and
CF = Cash Flow NBCR to justify the investment.
CRF = Capital Recovery Factor Using the equation of CF we get,
CC = Cost of Capital
CF = 21×200×10 = Rs.42,000
n = Number of years
I = Initial Investment Required From the table of interest factors [15], we find
that CRF = 6.1446 for a cost of capital of 10%
The values of CRF for a particular CC and n are (CC) and 10 years (n). Using the equations of
obtained from the table of interest factors [7]. NPV, BCR and NBCR, we get NPV= -156926,
Benefit Cost Ratio measures the present BCR = 0.622 and NBCR = -0.378. Since, NPV is
value/rupee of outlay, it is considered to be a negative, BCR less than one and NBCR is less than
useful criterion for ranking a set of alternatives in zero, duplicating machine 5 is not justified. It can
the decreasing use of capital. The Benefit cost ratio also be recalled that the machine 5 did not qualify
or the profitability index is given by for duplication based on cost workload index
criteria.
BCR = CF × (CRF, CC, n ) I (8) Repeating the same procedure for machine 11
(cash flow = Rs.42,000), we get NPV = 780732.2,
A variant of the benefit cost ratio is the Net BCR = 1.434 and NBCR = 0.434. Since, NPV is
Benefit Cost Ratio and is given by positive, BCR greater than one and NBCR greater
than zero, duplication of machine 11 is justified.
NBCR = NPV I (9) Considering machine 2, we have 3 machines of

82 - Vol. 17, No. 1, February 2004 IJE Transactions A: Basics

 6 cm 7 cm 8 cm 9 cm 10 cm 11 cm

Cell 4
Cell 1

13 16
Commo
8
n 6
4 3
14 17 9 5
11
2 7 19
1 18

Passage Passage Passage

5 cm
1 cm 2 cm 3 cm 4 cm

8 11
17

Cell3 13 18 15

20 12 10
Commo
n Cell 2
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8
1 cm 2 cm 3 cm 4 cm 5cm 6 cm 7 cm 8 cm
Scale X-Axis: 1cm = 5m Each cell is 0.25cm X 0.25cm
Scale Y-Axis: 1cm = 2.5m

this type. It belongs to cell 1 (Refer Figure 4b). It (13×200×10), NPV = -90240.4, BCR = 0.64 and
is a bottleneck machine as it is having intercell NBCR = -0.361 and hence this option is also not
movements with cells 2, 3 and 4. We can consider justified.
assigning of these additional machines to cells 2 The same procedure has been followed for all
and 3. bottleneck machines and machines, which
By duplicating machine 2 in cell 2 only one qualified on the basis of workload index criteria.
intercellular movement is reduced (Figure 4b). Alternatives that justified the investment include
Hence, its cash flow will be only Rs.2000 duplicating machine 11 in cell 4 and machines 8,
(1×200×10). Using the equations of NPV, BCR 13 in cell 1. Total costs of all these machines are
and NBCR we get, NPV = -237710.8, BCR = 0.05 within the budget constraint put by the company
and NBCR = -0.951. Hence, assigning machine 2 (Rs. 4 lacs). Forming remainder cell (cell dedicated
in cell 2 is not justified. Considering assigning to the manufacture of exceptional components
machine 2 to cell 3 we get CF = Rs.26,000 only) is also considered for processing exceptional

IJE Transactions A: Basics Vol. 17, No. 1, February 2004 - 83

components. Machines considered for the or though authors claim that their methods have
formation of remainder cell were 1,2,5,7,8,9,11,14 been applied successfully, details of their
and 15. Repeating the same procedure we get NPV implementation are not disclosed. Formation
= 1156326, BCR = 0.3946 and NBCR = -0.605. of machine cells, part family, processing of
Hence, this option is not feasible. exceptional components and designing of layout is
being dealt systematically. In most of the literature
Step 8 Arrange cells and machines within the available authors treat the problem of Cellular
cell: For arranging cells and for arranging Manufacturing as the problem of formation of
machines within the cell we have used from/to machine cells and part families. This paper
chart analysis. This includes the following steps: presents an integrated approach for cell
formation, processing of exceptional components
(a) Develop the from-to-chart from part routing and arrangement of Cellular layout. The paper
data. Machines 17, 18 (bench and layout) presented considers workload on each machine and
facilities are required commonly by all the cost before duplication. A total reduction of about
machines and hence are not used for from-to- 58% could be achieved in material handling by
chart calculation. implementing the proposed method. Other benefits
(b) Determine the to/from (T/F) ratio for each of Cellular Manufacturing such as reduced work-
machine (not shown). in-progress inventory, setup time etc. can be
Arrange cells/machines in order of increasing realized after the implementation of Cellular
to/from ratio (not shown). The notion is that Manufacturing.
cells/machines that have a low to/from ratio
receive work from other cells/machines but
distribute work to other cells/machines. Cells and
machines within the cells were arranged in 5. REFERENCES
accordance with to/from ratio obtained. Final
1. Srinivasan, G., “A Clustering Algorithm for Machine
layout obtained by the proposed method is shown Cell Formation in Group Technology Using Minimum
in Figure 5. Spanning Trees”, International Journal for Production
Material movement when traced through string Research, Vol. 32, No. 9, (1994), 2149-2158.
diagram revealed that there would be drastic 2. Crama, Y. and Oosten, “Models for Machine-Part
reduction in material handling, if the components Grouping in Cellular Manufacturing”, International
Journal for Production Research, Vol. 34, No. 6,
were processed through Cellular layout. With the (1996), 1693-1714.
present layout total material movement is 3148.2 3. Jayakrishnan, N. G. and Narendran, T. T., “On the
meters and through proposed layout it was found to Use of the Asymptotic Forms of the Boolean
be 1836.5 meters. Matrix for Designing Cellular Manufacturing
Systems- an Improved Approach”, European Journal
of Operations Research, Vol. 100, (1997), 429-
440.
4. McAuley, J., “Machine Grouping for Efficient
4. CONCLUSIONS Production”, Production Engineer, (1972), 53-
57.
The paper presented is an integrated approach for 5. Waghodekar, P. H. and Sahu, S., “Machine Component
Cell Formation in Group Technology: MACE”,
cell design. Cell formation dealt in this paper is International Journal for Production Research, Vol.
progressive i.e. it considers non-critical, ideal 22, No. 6, (1984), 937-948.
and critical machines individually. In real life 6. Parashar, N. and S., “A Heuristic Approach to Machine
situations this approach increases the scope for Cell and Part Family Formation in Group Technology”,
efficient machine cell formation and reduces the Industrial Engineering Journal, Vol. 24, No. 2,
(February 2000), 2-7.
complexity of the problem. Though many cell 7. Yashaswi, N. J., “Finance and Profits-The Managers
formation techniques are available in the literature Hand Book”, First Edition, Vision Books Pvt. Ltd, New
either they are not tested for practical applications Delhi, India, (1994).

84 - Vol. 17, No. 1, February 2004 IJE Transactions A: Basics