Documente Academic
Documente Profesional
Documente Cultură
Scatter search algorithm for supplier selection and order lot sizing under
multiple price discount environment
R. Mohammad Ebrahim a, J. Razmi a,*, H. Haleh b
a
b
a r t i c l e
i n f o
Article history:
Received 19 August 2008
Received in revised form 20 November 2008
Accepted 16 February 2009
Available online 26 March 2009
Keywords:
Supplier selection
Lot sizing
AHP
Price discount
MCDM
Scatter search
a b s t r a c t
Supplier selection is one the most important elements of supply chain management. This function
involves evaluation of many factors such as, cost of parts/materials, size of order, quality, and delivery
performance. Therefore, this problem is categorized as multi-criteria decision making problems. Different
approaches have been applied in order to assess and select the suppliers when suppliers offer discount on
the unit price. In practical conditions, buyers may face a situation where different types of discount may
be offered by candidate suppliers. None of the previous studies have considered different discount
schemes simultaneously. In this article a mathematical model is introduced which consider different
types of discount (all-unit cost, incremental discount, and total business volume discount) through
multi-objective formulation for single item purchasing problem. In addition, constraints such as suppliers capacity and demand are taken into consideration in the model. Due to the complexity of the problem
a proposed scatter search algorithm (SSA) is presented to solve this problem. Finally several sample problems have been solved by the proposed SSA and the exact (branch and bound) method. The results illustrate slight relative errors to compare with reasonable saving in computational times.
2009 Elsevier Ltd. All rights reserved.
1. Introduction
Supplier selection is one the most important components of
production and logistics management in the competitive environment of the global market. As organizations become more dependent on their suppliers the direct and indirect consequences of
poor decisions become more severe [4]. Such decisions involve
the selection of individual suppliers to employ and the determination of order quantities to be placed with the selected suppliers.
Practitioners need to follow strategies to achieve higher quality, reduced costs and shorter lead times to be able to compete in the global market. Within new strategies for purchasing and
manufacturing, suppliers play a key role in achieving corporate
competition [1,9,27].
Many attributes affect a suppliers performance. Dickson [8]
identied 23 criteria that have been considered by purchasing
managers in various supplier selection problems. Also Weber
et al. [26] found that 47 out of 76 articles which they reviewed, addressed more than one criterion for supplier selection decision
making. Hence supplier selection problem is a multiple criteria
decision making (MCDM) problem and it is necessary to make a
2. Literature review
Relevant studies in the literature of the subject have been divided into two major categories. First category of studies includes
the papers which consider just one criterion, usually the cost of the
procurement for supplier selection problem while suppliers offer
discounts on the quantity of materials being purchased.
Goossens et al. [13] discussed the procurement problem of buying multiple items from a set of suppliers considering just the minimization of purchasing cost function so other affecting criteria are
not included in their linear mathematical model. In their study
authors assume that all suppliers offer just all-unit discount scheme.
They argue that not only this problem is NP-hard but also there exists no polynomial-time approximation algorithm with a constant
ratio. Burke et al. [5] analyze the impact of supplier pricing schemes
and supplier capacity limitations on the optimal sourcing policy for
a single buyer. They consider the situation where the total quantity
to be procured for a single period is known by the rm and communicated to all suppliers. Each supplier quotes a price and capacity
limit as a maximum quantity that can supply to the buyer. According to this information, the buyer makes a decision for quantity allocation among the suppliers and consequently a subset of suppliers is
selected for order allocation. A variety of supplier pricing schemes
from the constituent group of suppliers is analyzed. Kothari et al.
[14] describe procurement auction with marginal decreasing piece
wise-constant supply curves. All-unit discount is allowed by this
auction. They present fully polynomial-time approximation
schemes for the winner determination problem and the computation of the corresponding payments of this auction. Benton [2] uses
Lagrangian relaxation to evaluate a purchasing managers resource
constrained order quantity decisions given alternative pricing
schedules from multiple suppliers. The author assumes that the
decision maker has a limited budget and storage space for ten items
offered by three suppliers, each quoting three all-units discount
intervals for each item. The objective is to minimize total acquisition
and inventory costs. The manager must choose a single supplier for
all items. Crama et al. [6] describe the purchasing decisions faced by
a multi-plant company. The suppliers of this company offer discount
schedules based on the total quantity of materials purchased. The
schedules simultaneously account both for corporate purchases
and for purchases at the individual plant level. The complexity of
the purchasing decisions is further increased due to the existence
of alternative production recipes for each nal product. They formulate the corresponding cost-minimization problem as a non-linear
mixed 01 programming problem. The above articles study the purchasing problem under only one discount offer by the suppliers. In
addition, their models consider just one objective (usually purchasing cost) which cannot be applied for the real cases.
767
Considering the importance of criteria such as quality of products and services provided by suppliers and also direct and indirect
impact of suppliers performance on organizations performance,
persuade organizations to consider other affective criteria in suppliers evaluation as well as cost. So the second category of papers
discussing supplier selection under discount environment declares
that supplier selection problem is a multi-objective decision making problem.
Rosenthal et al. [20] developed a mixed integer programming
model for supplier selection with bundling, in which a buyer needs
to buy various items from several vendors with limited capacity
and also with different quality and delivery performances which
offer bundled products at discounted prices. They applied single
objective programming in their model. Ghodsypour and OBrien
[11] proposed an integrated AHP and linear programming model
to help managers consider both qualitative and quantitative factors
in their purchasing activity in a systematic approach. In their model, price of the product has been offered in a discount scheme and
buyers limitations on budget, quality, and service have been taken
into account.
Amid et al. [1] consider a weighted additive fuzzy multi-objective model for the supplier selection problem under all-unit price
discounts. They formulate the problem in such a way as to simultaneously consider the imprecision of information and determine
the order quantities to each supplier based on price breaks. The
problem includes the three objective functions: minimizing the
cost, minimizing the rejected items and late deliveries, while satisfying capacity and demand requirement constraints. They argue, in
practice, for supplier selection problems, most of the input information is not known precisely, so they use the fuzzy optimization
theory to deal with this vagueness. A fuzzy weighted additive and
mixed integer linear programming is developed.
Dahel [7] presents a multi-objective mixed integer programming approach to simultaneously determine the number of suppliers to employ and the order quantities to allocate to these
suppliers in a multi-product, multi-supplier competitive sourcing
environment. The objectives to be optimized are: cost, delivery
and quality subject to the capacity constraints of suppliers. In their
selection problem the discount on the total business volume is
considered. Similar to the study of Dahel [7], Xia and Wu [28] discuss the multi-objective supplier selection problem under total
business volume discount environment. They propose an integrated approach of analytical hierarchy process improved by rough
sets theory and multi-objective mixed integer programming to
simultaneously determine the number of suppliers to employ
and the order quantity allocated to these suppliers while purchasing multiple products from a set of suppliers with multiple criteria.
In this study, they assume that suppliers offer price discount on total business volume. The above papers consider multiple objective
problems; however, their model cannot deal with multiple discount schemes simultaneously.
It is obvious that in practical situation buyers face with multiple
sourcing in which suppliers may offer different price discount
schemes. As it has been described above, there is need to develop
a model to consider various discount schemes simultaneously. In
this paper we intend to develop a mathematical model to consider
three discount schemes (all-unit discount, incremental discount
and total business volume discount) and some affecting factors
in a multi-objective, single-item supplier selection problem.
768
discussed the importance of criteria such as quality, on-time delivery and improvement of the performance of products in supplier
selection problem. Min and Galle [17] proposed seven criteria such
that each criterion has some sub-criteria.
By attention to the importance of intangible factors in evaluation of suppliers, a trade off between tangible and intangible factors should be made. Hence, we utilize some MADM methods
such as analytic hierarchy process (AHP) [22] to obtain a total
weight (accounted by AHP) for each supplier which implies that
the total performance of that supplier in intangible factors. So,
one of our objectives in this study is to maximize the total
weighted quantity of purchasing. We will discuss this objective later in this section in detail. We also consider three other objectives
as follows: minimizing the total cost of purchasing under discount
environment, minimizing the total defective items and minimizing
the total number of late delivered items. In the following sub-sections, objective functions and constraints are discussed in detail.
3.1. Assumptions and notations
The following assumptions are considered in denition and the
modeling of the problem.
The buyer can purchase required quantity from multiple
suppliers.
The buyer is going to purchase only one item from suppliers.
Each supplier can offer one of three discount schemes arbitrary.
The total demand is known and the total purchased quantity is
exactly equal to the total demand.
The total demand and purchased quantities from suppliers are
integer.
The buyer is purchasing just for one period.
Notation
index of suppliers
index of discount intervals
purchased quantity from supplier i in discount interval k
(integer variable)
binary variable; if the purchased quantity from supplier i
yik
falls on the interval corresponding to this variable then
yik = 1, otherwise yik = 0
n
total number of suppliers
number of suppliers that offer all-unit discount
n1
number of suppliers that offer incremental discount
n2
lower bound of the discount interval k offered by supplier i
lik
upper bound of the discount interval k offered by supplier i
uik
discounted unit price of the discount interval k offered by
pik
supplier i
defective rate of purchased goods from supplier i
di
late delivery rate of goods purchased from supplier i
hi
total weight of supplier i calculated by AHP
wi
index of the last interval offered by supplier i
Ki
D
total demand
i
k
xik
j1
XX
The total cost of purchasing depends only on the unit price of
item. The price for commodity items usually offered in different
discount schemes which can be seen in one of the following structures: All-unit discount, Incremental discount and total business
volume discount. In this section we propose a model in which all
of these discount models are considered simultaneously.
pik xik ;
i1 k1
djr V jr ;
j2Si r2Rj
where
X
r2Rj
V jr
pij xij ;
j 2 Si ;
i2C j
769
Vjr lies. With attention to Vj = pjxj which implies that the business
volume purchased from supplier j is the total quantity purchased
multiplied by the unit price without considering discount, and this
fact that Vj lies in an interval (for example r), so we have
V jr xj pj
ajr 6 V jr 6 bjr
or ljr 6 xj 6 ujr :
So, xj lies in an interval which depends on r so we can have the following relation:
XX
j2S
XX
djr pj xjr
XX
r2Rj
j2S
djr V jr
r2Rj
XX
j2S
r2Rj
j2S
XX
j2S
djr pj xjr
r2Rj
XX
j2S
djr V jr
r2Rj
pjr xjr :
r2Rj
where pjr is the discounted unit price of interval r offered by supplier j. This model is as same as all-unit discount model. h
So, in general, we just include all-unit and incremental discount
schemes in formulating the total cost function.
3.2.4. Total cost function
Let n1 suppliers offer all-unit discount schedules and n2 = n n1
suppliers offer incremental discount schedules. So by adding (1),
(2) we have
Ki
n1 X
X
pik xik
i1 k1
Ki
n
X
X
k1
X
in1 1 k1
Ki
n X
X
!
pij uij ui;j1
j1
Ki
n
X
X
xik pik
i1 k1
yik
k1
X
in1 1 k1
j1
min
Z1
Ki
n X
X
i1 k1
Ki
n
X
X
xik pik
in1 1 k1
k1
X
yik
!
j1
max
Z4
min
Z2
Ki
X
yik
k1
k1
X
di xik :
In order to reduce the total lead time of manufacturing, the number of production breakdowns, and the holding cost, buyers desire
to minimize the total late delivered items purchased from suppliers. So, the third objective function has been dened as Eq. (5):
Ki
n X
X
i1 k1
in1 1
j1
Criteria
Sub-criteria
Services
Background of relationships
Organizational level
Technological level
Level of information technology
Capital of the supplier
Flexibility in manufacturing
Communication
Z3
Table 1
Proposed criteria for supplier selection decision making.
i1 k1
min
min
Ki
n X
X
wi xik :
i1 k1
Ki
n X
X
hi xik :
770
min
Ki
X
n
X
aik xik
i1 k1
Ki
X
bik yik ;
di
hi
minr fwr g
;
W3
W4
wi
maxr fdr g
maxr fhr g
P
s1
bik W 1
j1 pij
maxr;s ; prs ;
P
yik 2 f0; 1g 8i 1; . . . ; n 8k 1; . . . ; K i
s1
j1 prj urj
pik ui;k1
o:
ur;j1 prs ur;s1
uij ui;j1
3.7. Constraints
There are some constraints associated with the supplier selection problem. In the following these constraints are explained
and modeled.
3.7.1. Demand constraint
The rst constraint with which we are faced is the demand constraint that implies the total purchased quantity must be equal to
the total demand of the buyer. This is modeled as follows:
Ki
n X
X
yik 6 1 8i 1; . . . ; n
xik 2 z [ f0g 8i 1; . . . ; n 8k 1; . . . ; K i
pik
P
n
o
s1
maxr;s prs ;
p
u
rj ur;j1 prs ur;s1
j1 rj
W2
8i 1; . . . ; n 8k 1; . . . ; K i
k1
in1 1 k1
where
aik W 1
Ki
X
xik D:
10
i1 k1
8i 1; . . . ; n 8k 1; . . . ; K i ;
yik 1 8i 1; . . . ; n:
11
12
k1
xik 6 C i
8i 1; . . . ; n:
aik xik
i1 k1
s:t:
Ki
n X
X
i1 k1
Ki
n
X
X
in1 1 k1
xik D
k1
min
bik yik
1.
2.
3.
4.
5.
6.
In the process of equalizing the intervals it happens for suppliers with lower capacities that all of their offered intervals are
equalized before Algorithm 1 is stopped. In this case, we omit
the index of the corresponding supplier from the set of suppliers
for rest of process. Fig. 2 illustrates this algorithm.
Vector solutions in this problem have the following general
format:
1
S1
Sn
z}|{
z}|{
S @x11 ; . . . ; x1K 1 ; . . . ; xn1 ; . . . ; xnK n A
771
13
4.2.2.2. Phase 2. In this phase, for each vector solution, those assigned quantities (positive quantity variables) which are lower
than a xed determined quantity are considered. This xed quantity which is the lowest acceptable quantity that must be pur-
772
v
u n Kr
uX X
distij t
xirk xjrk 2 ;
r1 k1
Begin
Initializing (Set P):
for (q:=1) to (q:=200) do{
p:=q (mod) n;
if (p=0) then p:=n;
for (p:=1) to (p:=n) do
{ xp:=random (1, min{ capacity of sup. p & D});
D:=D- xp;
p:= p+1;} While( D>0)
Determine the corresponding interval for x p;}
// end of Initializing (Set P);
Impv. Method:
Phase 1:
for (i:=1) to (i:=n-1) do
{ for (j:=i+1) to (j:=n) do {
if (capacity of sup. i >=xj & capacity of sup. j>=xi) then (swap the values of x i and
xj);
if ( func. value( new)< func. value( old)) then (keep this swapping fixed)
otherwise (discard it); }
Phase 2:
Determine a minimum order limit (MOL);
sum:=0;
for (i:=1) to (i:=n) do {
if (xi <= MOL) then ( sum:=sum+ xi);}
for (i:=1) to (i:=n) do { allocate sum to sup. i if possible;
if ( func. value( new)< func. value( old)) then (keep this allocation fixed)
otherwise (discard it);
// end of Impv. Method;
Ref-Set Updating:
for (i:=1) to (i:=maximum index of solutions available) do
{ sort according to Func. Value)}
for (j:=1) to (j:=10) do
{ copy the jth sorted solution in Ref-Set) }
for (i:=11) to (i:=maximum index of solutions available) do
{ compute dist =
ij
Kr
( x
r =1 k =1
i
rk
x rkj )
dist i* = m in {dist i j }
j
773
where distij is the distance of solution i from solution j. Now the distance index is measured as follows:
dist i minfdist ij g:
j
The number of b2 solutions with highest distance index value is selected. These are sorted with respect to their distance and placed in
the second subset.
4.2.4. Subset generation method (constructing new solution sets)
In this step, new subsets are constructed from Ref-Set. Let the
Ref-Set have b1 + b2 members, where b1 is the number of members
in the rst section of Ref-Set and b2 is the number of members in
the second section of Ref-Set containing diverse solutions. Therefore, there are 2b1 b2 1 non-empty subsets that can be constructed from Ref-Set. Depend on the combination method; some
of these subsets may be useful. In this paper, in order to optimize
the computational time, we use only the subsets which have just
two members.
4.2.5. Solution combination method
The inputs of this step are outputs of the previous step. Members of the subsets made in the previous step are combined together and consequently new solutions are obtained. In this
paper, the convex combination of each couple of solutions which
are the members of subsets constructed in the previous step is
computed. Since the value of k (parameter of convex combination)
is considered to be 0.5, the outputs are feasible.
In our proposed SSA these ve steps are utilized as shown in
Fig. 3. In the following section some numerical examples are provided to illustrate how the problem formulation and solutions
can be altered by different situations.
min
s:t:
Table 2
Discount intervals of example.
# of supplier
Intervals
Unit prices
1100
101200
201300
301400
623
534
465
383
150
51100
101150
515200
201250
251300
301350
351400
401450
451500
670
610
560
510
470
440
400
360
320
290
180
81140
141200
201280
281350
351400
401500
660
612
550
520
450
370
280
175
76150
151225
226300
301375
650
620
530
460
390
Rate of defective
items (di)
Final weight of
suppliers (wi)
1
2
3
4
0.009
0.02
0.008
0.006
0.01
0.009
0.01
0.008
0.2016
0.3222
0.2734
0.2028
774
Table 4
Matrix of pair wise comparison ratios of criteria of Table 1.
Criterion i
Services
Background of
relationships
Organizational
level
Communication
Criterion j
Services
Background of
relationships
Organizational
level
Communication
1
1/4
4
1
2
1/2
7
1/3
1/2
1/2
1/7
sequently, involving multiple criteria make the prediction of optimal policy and opted suppliers difcult. In addition, in case of
regarding just the cost factor, both discount model and discount
percentages are inuencing the problem and its optimal solution.
Example 3. In this example we intend validate our developed linear integer mathematical programming by comparing it with the
non-linear mathematical programming model described by Burk
et al. [5]. In their model they have assumed that the only criterion
is the cost of purchasing and in addition suppliers are allowed to
offer just one type of discount such as incremental discount.
Therefore, given the data of Example 1, suppose that similar to
Example 2 just the cost function is considered. More over, suppliers
are assumed to offer just incremental discount. This problem is
modeled using both model of Burk et al. and the developed model
of this paper. Running LINGO with B&B solver method, it is seen
that optimal solutions obtained by solving both models are the
same which is purchasing the whole 500 units of demand from
supplier 2 in discount interval 16. Therefore, x2,16 = 500, y2,16 = 1
and other variables are zero. The computational time of B&B for
the model of Burk et al has been 5 min and 33 s while this time
consumed by B&B for solving our model has been 2 s. This problem
is solved by the SSA as well and the same outputs have been
obtained within just 0.7 s.
Table 5
Objective coefcients and upper and lower bounds of discount intervals.
Number of interval
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplier 1
Supplier 2
Supplier 3
Supplier 4
aik
bik
aik
bik
aik
bik
aik
bik
2.456189
2.456189
2.456189
2.456189
2.455229
2.455229
2.455229
2.454553
2.454553
2.454553
2.454553
2.45375
2.45375
2.45375
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2.532259
2.531671
2.531671
2.531671
2.531182
2.531182
2.530692
2.5303
2.5303
2.530007
2.530007
2.529615
2.529223
2.529223
2.528832
2.528538
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2.143844
2.143844
2.143844
2.143374
2.143374
2.142767
2.142767
2.142473
2.142473
2.142473
2.141788
2.141788
2.141004
2.141004
2.140123
2.140123
0
0
0
0.037
0.037
0.122
0.122
0.181
0.181
0.181
0.373
0.373
0.647
0.647
1
1
2.100448
2.100448
2.100154
2.100154
2.100154
2.100154
2.099273
2.099273
2.098588
2.098588
2.098588
2.097902
2.097902
0
0
0.022
0.022
0.022
0.022
0.154
0.154
0.308
0.308
0.308
0.514
0.514
1
51
76
81
101
141
151
201
226
251
281
301
351
376
401
451
50
75
80
100
140
150
200
225
250
280
300
350
375
400
450
500
775
n1, n2
Relative
Error
SSA computational
time
n1 = 4, n2 = 0
n1 = 0, n2 = 4
n1 = 2, n2 = 2
1277.646
1055.241
1055.186
1277.646
1055.241
1055.186
0
0
0
0.4
0.7
0.4
1.3
3.7
1.0
n1 = 8, n2 = 0
n1 = 0, n2 = 8
n1 = 4, n2 = 4
692.30
345.728
345.728
692.30
345.728
345.728
0
0
0
0.6
0.7
0.5
1.2
3.1
0.9
12
n1 = 12,
n2 = 0
n1 = 0,
n2 = 12
n1 = 6, n2 = 6
1287.7
0.4
0.5
2.3
0.3
2.5
n1 = 15,
n2 = 0
n1 = 0,
n2 = 15
n1 = 8, n2 = 7
1467.0
1467.0
0.6
2.4
1196.87
1196.87
0.5
3.5
1195.24
1195.24
0.6
2.9
1440.0
1440.0
0.4
3.3
1266.93
1266.93
0.5
3.8
1266.93
1266.93
0.7
5.2
n1 = 25,
n2 = 0
n1 = 0,
n2 = 25
n1 = 15,
n2 = 10
1471.026
1471.026
0.5
4.1
1182.496
1182.496
1.2
3.7
1181.671
1181.671
0.6
5.4
n1 = 30,
n2 = 0
n1 = 0,
n2 = 30
n1 = 15,
n2 = 15
1512.68
1500.0
0.0084
0.8
6.2
1443.05
1485.93
0.0297
1.5
7.6
1443.05
1485.93
0.0297
1.8
7.1
n1 = 40,
n2 = 0
n1 = 0,
n2 = 40
n1 = 20,
n2 = 20
1893.7
1887.53
0.0033
1.3
9.3
1497.128
1497.128
0.0
1.5
12.2
1497.46
1524.052
0.0177
1.1
10.8
15
20
25
30
40
n1 = 20,
n2 = 0
n1 = 0,
n2 = 20
n1 = 10,
n2 = 10
974.694
974.15
1287.7
974.694
974.15
14
The mean value of these relative errors is 0.0037 and the variance is 7.95809E05. In addition mean computational time of
the SSA is 0.754 and the corresponding variance is 0.1722 that
shows the SSA is adequately fast in nding solutions. Therefore,
we can claim that the SSA is largely reliable and can be regarded
as a suitable alternative for exact algorithms such as B&B in solving
the described supplier selection problem.
7. Summary and conclusions
In supplier selection decision making there are tangible and
intangible conicting factors that affect this decision making problem. Furthermore, in many real situations, suppliers may offer different discount schemes according to their selling policies. Hence,
buyers face a complex decision making problem. This paper intro-
LINGO computational
time
776
References
[1] Amid A, Ghodsypour SH, OBrien Ch. A weighted additive fuzzy multi-objective
model for the supplier selection problem under price breaks in a supply Chain.
Int J Prod Econom 2007. doi:10.1016/j.ijpe.2007.02.04.
[2] Benton WC. Quantity discount decisions under conditions of multiple items,
multiple suppliers and resource limitations. Int J Prod Res 1991;29(10):195361.
[3] Benton WC, Park SA. Classication of literature on determining the lot size
under quantity discount. Eur J Oper Res 1996;92(2):21938.
[4] Boer L, Labro E, Morlacchi P. A review of methods supporting supplier
selection. Eur J Purchasing Supply Manage 2001;7:7589.
[5] Burke GJ, Carrillo J, Vakharia AJ. Heuristics for sourcing from multiple suppliers
with alternative quantity discounts. Eur J Oper Res 2008;186(1):31729.
[6] Crama Y, Pascual R, Torres A. Optimal procurement decisions in the presence of
total quantity discounts and alternative product recipes. Eur J Oper Res
2004;159:36478.
[7] Dahel N-E. Vendor selection and order quantity allocation in volume discount
environments. Supply Chain Manage 2003;8(4):33542.
[8] Dickson GW. An analysis of vendor selection systems and decisions. J
Purchasing 1966;2(1):517.
[9] Dobler DW et al. Purchasing and materials management: text and cases. New
York: McGraw-Hill; 1990.
[10] Dolan RJ. Quantity discounts managerial issues and research opportunities.
Market Sci 1987;6(l):122.
[11] Ghodsypour SH, OBrien C. An integrated method using the analytical
hierarchy process with goal programming for multiple sourcing with
discounted prices. In: Proceedings of 14th international conference on
production research (ICPR), Osaka, Japan; 1997.
[12] Glover F. Heuristics for integer programming using surrogate constraints.
Decision Sci 1977;8:15666.
[13] Goossens DR, Maas AJ T, Spieksma FCR, Klundert JJ. Exact algorithms for
procurement problems under a total quantity discount structure. Eur J Oper
Res 2007;178:60326.
[14] Kothari A, Parkes D, Suri S. Approximately-strategy proof and tractable multiunit auctions. In: Proceedings 4th ACM conference on electronic commerce
(EC-2003), San Diego, June 2003. ACM, p. 16675.
[15] Mart R. Scatter searchwellsprings and challenges. Eur J Oper Res
2006;169:3518.
[16] Mart R, Languna M, Glover F. Principles of scatter search. Eur J Oper Res
2006;169:35972.
[17] Min H, Galle WP. International supplier selection. Int J Phys Distrib Log
Manage 1994;24(5):2433.
[18] Nahmias S. Production and operations analysis. 4th ed. New York: McGraw/
Irwin; 2001.
[19] Rahimi-Vahed AR, Rabbani M, Tavakkolli Moghaddam R, Torabi SA, Jolai F. A
multi-objective scatter search for a mixed-model assembly line sequencing
problem. Adv Eng Inform 2007;21:8599.
[20] Rosenthal EC, Zydiac JL, Chaudhry SS. Vendor selection with bundling. Decision
Sci 1995;26(1):3548.
[21] Rubin PA, Benton WC. Evaluating jointly constrained order quantity
complexities for incremental discounts. Eur J Oper Res 2003;149:
55770.
[22] Saaty TL. The analytic hierarchy process. NY: McGraw-Hill; 1980.
[23] Shih H-Sh, Shyur H-J, Lee ES. An extension of TOPSIS for group decision
making. Math Comput Model 2007;45:80113.
[24] Tracey M. Empirical analysis of supplier selection and involvement. Supply
Chain Manage: Int J 2001;6(4):17488.
[25] Wang G, Huang SH, Dismukes JP. Product-driven supply chain selection using
integrated multi-criteria decision making methodology. Int J Prod Econ
2004;91:115.
[26] Weber CA, Current JR, Benton WC. Vendor selection criteria and methods. Eur J
Oper Res 1991;50:218.
[27] Willis HT, Huston RC, Pohlkamp F. Evaluation of just in time supplier
performance. Prod Inventory Manage J 1993;34(2):15.
[28] Xia W, Wu Zh. Supplier selection with multiple criteria in volume discount
environments. Omega 2007;35:4945.