Multiple Criteria ABC + AHP

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Computers & Operations Research 33 (2006) 695 700

www.elsevier.com/locate/cor

ABC inventory classication with multiple-criteria using


weighted linear optimization
Ramakrishnan Ramanathan
Operations Management and Business Statistics, College of Commerce and Economics, Sultan Qaboos University, Post Box
20, Postal Code 123, Sultanate of Oman, Oman
Available online 24 August 2004

Abstract
Inventory classication using ABC analysis is one of the most widely employed techniques in organizations.
The need to consider multiple criteria for inventory classication has been stressed in the literature. A simple
classication scheme is proposed in this paper using weighed linear optimization. The methodology is illustrated
using an example.
2004 Elsevier Ltd. All rights reserved.

Keywords: ABC Inventory classication; Multiple criteria; Weighted linear optimization

1. Introduction

Inventory classication using ABC analysis is one of the most widely employed techniques in orga-
nizations. This classication is based on the Pareto principle. ABC analysis is easy to use and simple to
understand by an average materials manager. Normally, the items are classied based on the annual use
value, which is the product of annual demand and average unit price. Class A items are relatively few in
number but constitute a relatively large amount of annual use value, while class C items are relatively
large in number but constitute a relatively small amount of annual use value. Items between the above
two classes constitute class B, though some studies claim that there is no need to include this class in

Tel.: +968-513-333x2849; fax: +968-514-043.


E-mail address: ramanathan@squ.edu.om (R. Ramanathan).

0305-0548/$ - see front matter 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.cor.2004.07.014
696 R. Ramanathan / Computers & Operations Research 33 (2006) 695 700

the analysis [1]. Class A items have to be controlled tightly and monitored closely. Textbooks such as
Silver et al. [2] provide more details on inventory control policies for these classes of items. ABC analysis
is successful only when the inventory being classied is fairly homogeneous and the main difference
among the items is in its annual use value (computed from unit price and demand volume). In practice, an
organization of even moderate size has to control thousands of inventory items and they need not be very
homogeneous. As more and more customers demand a wide range of products, the need to increase the
variety of inventory items is also increasing. Thus, it has been generally recognized that the traditional
ABC analysis may not be able to provide a good classication of inventory items in practice [35].
There are many instances when other criteria, other than the annual use value, become important [3]
in deciding the importance of an inventory item. This problem of multi-criteria inventory classication
(MCIC) has been addressed by some studies in the literature. Some of the criteria considered in the
literature include inventory cost, part criticality, lead time, commonality, obsolescence, substitutability,
number of requests for the item in a year, scarcity, durability, substitutability, repairability, order size
requirement, stockability, demand distribution, and stock-out penalty cost [1,3,58].
Complex computational tools are needed for multi-criteria ABC classication. Flores et al. [9] provide
a matrix-based methodology. A joint criteria matrix is developed in the case of two criteria. However,
the methodology is relatively difcult to use when more criteria have to be considered. Several multiple-
criteria decision-making (MCDM) tools have also been employed for the purpose. Cohen and Ernst
[10] and Ernst and Cohen [11] have used cluster analysis to group similar items. The analytic hierarchy
process (AHP) [17] has been employed in many MCIC studies [9,1214]. When AHP is used, the general
idea is to derive a single scalar measure of importance of inventory items by subjectively rating the
criteria and/or the inventory items [9,3]. The single most important issue associated with AHP-based
studies is the subjectivity involved in the analysis. Heuristic approaches based on articial intelligence,
such as genetic algorithms [3] and articial neural networks [5], have also been applied to address the
MCIC problem. Clearly, these approaches are heuristics and need not provide optimal solutions at all
environments. In this paper, we propose a simple weighted linear optimization model to address the MCIC
problem.

2. Model development

Assume that there are N inventory items, and that the items have to be classied as A, B or C based on
their performance in terms of J criteria. In particular, let the performance of mth inventory item in terms
of each of the criteria be denoted as ymj . Let us further assume that all the criteria are positively related to
the importance level of the itemi.e., the larger the score of an item in terms of these criteria, the greater
is the chance that the item be classied as an A-Class item. This assumption is made because most of
the criteria used in inventory classication [13,58] are positively related. Even if there are inversely
related criteria, reciprocals of the scores could be used to make them positive criteria.
In the proposed approach, a weighted additive function is used to aggregate the performance of an
inventory item in terms of different criteria to a single score, called the optimal inventory score of
an item. The weights are chosen using optimization subject to the constraints that the weighted sum,
computed using the same set of weights, for all the items must be less than or equal to one. The model
R. Ramanathan / Computers & Operations Research 33 (2006) 695 700 697

is shown below:

J
max vmj ymj ,
j =1
J
vmj ynj  1, n = 1, 2, . . . , N, (Model 1)
j =1
vmj  0, j = 1, 2, . . . , J.
The model uses a maximization objective function as all the criteria are assumed to be positively related
to the importance level of an item. The model, when solved, gives the optimal inventory score for the
mth item. To get the optimal scores of other inventory items, Model (1) should be solved repeatedly by
changing the objective function. These scores can then be used to classify the inventory items.
It is worth pointing out the similarities of Model (1) with a class of linear programming models used in
data envelopment analysis (DEA) [15]. An output maximizing multiplier DEA model with many outputs
and a constant input will reduce to Model (1).

2.1. An illustration

The weighted additive model is applied in this section for classifying inventory using the data provided
in Flores et al. [9]. The article used data on traditional ABC inventory classication presented in Reid [16]
and applied a multi-criteria classication methodology based on AHP. The ABC classication in Flores
et al. [9] was based on four criteria, namely, average unit cost, annual dollar usage, critical factor (1 for
a very critical item, 0.01 for a non-critical item and 0.50 for a moderately critical item), and lead time
(ranging from 1 to 7 weeks). The data are shown in Table 1. Note that all the four criteria are positively
related to the importance level of inventory items. As mentioned earlier, the optimal inventory scores for
all the 47 inventory items are computed using Model (1) repeatedly by changing the objective function
for each item. In order to get the same distribution of 10 class A items, 14 class B items and 23 class C
items, inventory items with optimal efciency scores of 86% and above are classied as class A items,
items with scores of 60% and below are classied as class C items, while items with scores between
86% and 60% are classied as class B items. Note that the classication using AHP is based on weights
of 0.079, 0.091, 0.420 and 0.410, respectively, for the four criteria [9]. Thus weights for the rst two
criteria (average unit cost and annual dollar usage) are much smaller than the weights for the remaining
two criteria (critical factor and lead time).
As shown in Table 1, the ABC classication using Model (1) provides different results compared
with the traditional (i.e., based on annual dollar usage) and AHP-based classication. Model (1) results
coincide with the other two classication schemes for about 40% of the items. The difference could be
because of the underlying assumption behind the methods. For example, consider Item S4. This item
is considered as a class A item based on annual dollar usage as it has one of the highest annual dollar
usage. It has been classied as a class A item by Model (1) as well, for the same reason. However, Item
S4 is classied as class C item by the AHP-based method because of the weighting scheme adopted in
the method (that provides lower weight for annual dollar usage). Item S14 is classied by Model (1) as a
class A item as it has one of the highest values in terms of average unit cost and lead time, but the same
item is classied as a class B item by the classication based on annual dollar usage as it has a moderate
value in terms of annual dollar usage. Item S14 is classied as a class B item by the AHP-based method
because of the nature of weighting scheme adopted.
698 R. Ramanathan / Computers & Operations Research 33 (2006) 695 700

Table 1
A comparison of ABC classication using optimal inventory score of the proposed model (Model 1), annual dollar usage, and
AHP methodologies

Item no. Average Annual Critical Lead Optimal ABC classication using
unit cost ($) dollar usage ($) factor time inventory score (%)

Optimal Annual AHP


inventory score dollar weighted
score usage score

S4 27.73 4769.56 0.01 1 100.0 A A C


S14 110.4 883.2 0.5 5 100.0 A B B
S18 49.5 594 0.5 6 100.0 A B A
S29 134.34 268.68 0.01 7 100.0 A C B
S34 7.07 190.89 0.01 7 100.0 A C C
S45 34.4 34.4 0.01 7 100.0 A C B
S15 71.2 854.4 1 3 96.8 A B A
S16 45 810 0.5 3 91.7 A B C
S28 78.4 313.6 0.01 6 89.0 A C C
S19 47.5 570 0.5 5 86.6 A B B
S40 51.68 103.36 0.01 6 85.7 B C C
S17 14.66 703.68 0.5 4 79.9 B B B
S31 72 216 0.5 5 72.4 B C B
S33 49.48 197.92 0.01 5 71.7 B C C
S23 86.5 432.5 1 4 71.5 B B A
S37 30 150 0.01 5 71.4 B C C
S39 59.6 119.2 0.01 5 71.4 B C C
S43 29.89 59.78 0.01 5 71.4 B C C
S47 8.46 25.38 0.01 5 71.4 B C C
S20 58.45 467.6 0.5 4 69.9 B B B
S21 24.4 463.6 1 4 69.7 B B A
S22 65 455 0.5 4 69.4 B B B
S27 84.03 336.12 0.01 1 67.1 B C C
S1 49.92 5840.64 1 2 61.9 B A A
S24 33.2 398.4 1 3 54.4 C B A
S26 33.84 338.4 0.01 3 51.8 C C C
S38 67.4 134.8 0.5 3 50.2 C C C
S35 60.6 181.8 0.01 3 46.7 C C C
S2 210 5670 1 5 45.1 C A A
S36 40.82 163.28 1 3 44.9 C C B
S30 56 224 0.01 1 44.7 C C C
S6 31.24 2936.67 0.5 3 42.9 C A C
S44 48.3 48.3 0.01 3 42.9 C C C
S46 28.8 28.8 0.01 3 42.9 C C C
S32 53.02 212.08 1 2 42.4 C C B
S25 37.05 370.5 0.01 1 41.9 C C C
S10 160.5 2407.5 0.5 4 40.4 C A B
S7 28.2 2820 0.5 3 37.8 C A C
S5 57.98 3478.8 0.5 3 30.8 C A B
R. Ramanathan / Computers & Operations Research 33 (2006) 695 700 699

Table 1 (continued)

Item no. Average Annual Critical Lead Optimal ABC classication using
unit cost ($) dollar usage ($) factor time inventory score (%)

Optimal Annual AHP


inventory score dollar weighted
score usage score

S41 19.8 79.2 0.01 2 28.7 C C C


S42 37.7 75.4 0.01 2 28.6 C C C
S8 55 2640 0.01 4 26.7 C A C
S9 73.44 2423.52 1 6 16.4 C A A
S11 5.12 1075.2 1 2 12.0 C B B
S13 86.5 1038 1 7 9.8 C B A
S3 23.76 5037.12 1 4 5.4 C A A
S12 20.87 1043.5 0.5 5 4.0 C B B
Source for the data and classications based on annual dollar usage and AHP: Ref. [9]

3. Summary

A weighted linear optimization model has been proposed and illustrated in this paper for classifying
inventory items in the presence of multiple criteria. It is a very simple model that can be easily understood
by inventory managers. The model is similar to linear programming models employed in data envelopment
analysis.

References

[1] Hautaniemi P, Pirttil T. The choice of replenishment policies in an MRP environment. International Journal of Production
Economics 1999;59:8592.
[2] Silver EA, Pyke DF, Peterson R. Inventory management and production planning and scheduling. New York: Wiley; 1998.
[3] Guvenir HA, Erel E. Multicriteria inventory classication using a genetic algorithm. European Journal of Operational
Research 1998;105:2937.
[4] Huiskonen J. Maintenance spare parts logistics: special characteristics and strategic choices. International Journal of
Production Economics 2001;71:12533.
[5] Partovi FY, Anandarajan M. Classifying inventory using an articial neural network approach. Computers and Industrial
Engineering 2002;41:389404.
[6] Duchessi P, Tayi GK, Levy JB. A conceptual approach for managing of spare parts. International Journal of Physical
Distribution and Materials Management 1988;18(5):815.
[7] Flores BE, Whybark DC. Multiple criteria ABC analysis. International Journal of Operations and Production Management
1986;6(3):3846.
[8] Stonebraker PW, Leong GK. Operations strategy: focusing competitive excellence. Allyn and Bacon: Boston; 1994.
[9] Flores BE, Olson DL, Dorai VK. Management of multicriteria inventory classication. Mathematical and Computer
Modeling 1992;16(12):7182.
[10] Cohen MA, Ernst R. Multi-item classication and generic inventory stock control policies. Production and Inventory
Management Journal 1988;29(3):68.
[11] Ernst R, Cohen MA. Operations related groups (ORGs): a clustering procedure for production/inventory systems. Journal
of Operations Management 1990;9(4):57498.
[12] Gajpal PP, Ganesh LS, Rajendran C. Criticality analysis of spare parts using the analytic hierarchy process. International
Journal of Production Economics 1994;35(13):2937.
700 R. Ramanathan / Computers & Operations Research 33 (2006) 695 700

[13] Partovi FY, Burton J. Using the analytic hierarchy process for ABC analysis. International Journal of Production and
Operations Management 1993;13(9):2944.
[14] Partovi FY, Hopton WE. The analytic hierarchy process as applied to two types of inventory problems. Production and
Inventory Management Journal 1993;35(1):139.
[15] Charnes A, Cooper WW, Lewin AY, Seiford LM. Data envelopment analysis: theory methodology and applications. Kluwer:
Boston; 1994.
[16] Reid RA. The ABC method in hospital inventory management: a practical approach. Production and Inventory Management
1987;28(4):6770.
[17] Saaty TL. The analytic hierarchy process. McGraw-Hill: New York; 1980.

You might also like