Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
(Received: February 27, 2016 / Revised: November 1, 2016 / Accepted: December 6, 2016)
ABSTRACT
This paper discusses a reverse supply chain (RSC) which consists of the process flows from procurement of used
products collected from a market, through remanufacturing products from the used products, to sales of the products
in a market. In general, it is conceivable for the RSC to face the uncertainty in quality of used products collected from
a market. Inspection is one of efficient methods to deal with the problem regarding quality of used products. However,
there is a trade-off between inspection cost and inspection accuracy. This paper focuses on the following five types of
inspection: (1) 100% inspection, (2) sampling inspection, (3) sampling inspection with screening of rejected lots, (4)
sampling inspection with screening of acceptable lots, and (5) no inspection, and determines the optimal operation
consisting of the optimal number of procured used products and the optimal inspection policy. Numerical analysis
clarifies not only how changes of conditions of the RSC affect the manufacturers optimal operation but also features
of each inspection type. In addition, from the results of numerical analysis, this paper shows the usability to add the
proposed inspection in this paper, the sampling inspection with screening of acceptable lots, to choices of inspection
type.
Keywords: Reverse Supply Chain, Uncertainty in Quality of Used Products, Sampling Inspection, Inspection Policy,
Optimal Operation
Korugan (2009), Ferguson et al. (2009), Fleischman et al. were determined by using the ARENA simulation pro-
(1997), Guide (2000), Guide and Wassenhove (2001), gram.
Inderfurth (2005), Konstantaras et al. (2010), Kusukawa Inspection is one of efficient methods to deal with
and Akizawa (2015), Nenes et al. (2010), Pochampally et the uncertainty in quality of used products regarding the
al. (2009), Pokharel and Liang (2012), Savaskan and Van remanufacturing. Theoretical analyses and the marginal
Wassenhove (2006), Tang and Teunter (2006), Teunter insights obtained from numerical examples regarding the
and Flapper (2011), Watanabe and Kusukawa (2014), effects of inspection and sorting of used products on the
Watanabe et al. (2013), Wei et al. (2011), Wu (2012), optimal tactical production planning in RSC are discussed
Zhang et al. (2014). in the following previous papers: Aras et al. (2004), Be-
Also, a supply chain which organizes the forward hret and Korugan (2009), Ferguson et al. (2009), Guide
chains and the reverse chains has been called a closed- and Van Wassenhove (2001), Guide et al. (2003), Kons-
supply chain or reverse supply chain, which is defined in tantaras et al. (2010), Nenes et al. (2010), Tagaras and
the following previous papers: Fleischman et al. (1997), Zikopoulos (2008), Zikopoulos and Tagaras, (2007, 2008).
Guide et al. (2003), Inderfurth (2005), Shi et al. (2010, Zikopoulos and Tagaras (2007, 2008) and Van Wassen-
2011), Tagaras and Zikopoulos (2008), Thierry et al. hove and Zikopoulos (2010) incorporated some quality
(1995), Van Wassenhove and Zikopoulos (2010), Wei et classification errors into a RSC. Here, quality classifica-
al. (2012), Yan and Sun (2012), Zhang et al. (2014), Zi- tion errors were dealt as an uncertain value. Zikopoulos
kopoulos and Tagaras (2007, 2008). and Tagaras (2008) handled quality classification errors as
In this study, a supply chain with the forward chains Type I error and Type II error based on a concept of quali-
and the reverse chains is referred to as a reverse supply ty control. However, in their previous studies, possible
chain (RSC). The concept and management of RSC has errors in inspection and sorting of used products were
been evolving to assist practitioners, academic research- considered as a given Type I error and a given Type II
ers and real-world policymakers regarding production error, and those errors were incorporated into the RSC.
planning in a RSC which try to take some measures and Besides, in the above previous studies, a concept of the
policies in order to promote 3R activities (Reduce-Reuse- optimal decision-making for the lower limit of quality to
Recycle) in the RSC. recycle used products was not considered for a RSC.
In general, it is conceivable for the RSC to face the However, above previous papers did not discussed a
uncertainty in quality of used products collected from a trade-off between inspection cost and inspection accuracy.
market. It is necessary for practitioners, academic re- When inspection processes are incorporated into opera-
searchers and real-world policymakers to consider the tions in a RSC, it is necessary for firms related to opera-
optimal operations for RSC, and the uncertainty in quality tion of the RSC to consider a trade-off between inspection
of used products regarding the remanufacturing. Several cost and inspection accuracy. In addition, it isnt always
previous papers have dealt with the optimal operations for possible to develop and apply an adequately accurate in-
RSC, and the uncertainty in quality in used products re- spection technology at low cost. In the cases, a sampling
garding the remanufacturing in a RSC has been attracting inspection may be a realistic alternative to mitigate the
more attention in recent papers. Aras et al. (2004) investi- inspection cost (Lee and Rosenblatt. 1985, Panagiotidou
gated the effect of the stochastic nature of product returns et al., 2013). Panagiotidou et al. (2013) discussed the
on the optimal production planning and verified condi- optimal decision for a procurement quantity of used
tions under which quality-based categorization was most products and a sampling size. A sampling inspection for a
cost effective. Zikopoulos and Tagaras (2007) investi- batch of used products was conducted before the pro-
gated how uncertainty regarding the quality of returned curement. The procurement quantity and the sample size
products in two collection sites affected the profitability were determined according to the estimated quality condi-
of reuse activities and derived the unique optimal solution tion from results of the sampling inspection. Establishing
(procurement and production quantities). In Guide and a sampling inspection might produce the potential eco-
Van Wassenhove (2001) and Ferguson et al. (2009), re- nomic benefits. However, this previous paper didnt suffi-
turned products were assumed to have N quality catego- ciently consider how disposal of un-remanufactured used
ries, and the procurement prices and the remanufacturing products affected the expected profit of a remanufacturer.
costs were different based on the corresponding quality Here, a sampling inspection has two essential demerits: (I)
level. Behret and Korugan (2009) discussed a remanufac- remanufacturable used products are disposed and (II)
turing stage with uncertainties in the quality of remanu- unremanufacturable used products are conveyed to a re-
facturing products, return rates and return times of re- manufacturing process.
turned products. After returned products were classified This paper focuses on the optimal operation for a
by considering the uncertainty in quality of the returned RSC to remanufacture a single type of used products such
product, remanufacturing processing times, material re- as consumer electronics (mobile phone, personal comput-
covery rates, the remanufacturing costs and disposal costs er), semiconductor and electronic component. (Guide,
Optimal Operation for Reverse Supply Chain Incorporating Inspection Policy into Remanufacturing of Used Products
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 3
2000; Guide and Jayaraman, 2000; Guide and Van Was- bility to add inspection types (4), which is the proposed
senhove, 2001; Guide et al., 2003; Ferguson et al., 2009; inspection in this study, to choices of inspection type for a
Van Wassenhove and Zikopoulos, 2010). When the RSC RSC.
is operated to remanufacture parts extracted from used Contribution of this paper is to consider the follow-
products, it is necessary for firms to discuss how the qual- ing topics regarding a RSC.
ity of used products collected from a market and a trade- y The 5 ITs: (IT-1) 100% inspection, (IT-2) sampling in-
off between inspection cost and inspection accuracy af- spection, (IT-3) sampling inspection with screening of
fects the optimal operation for a RSC. Also, firms may rejected lots, (IT-4) sampling inspection with screening
need to determine the optimal operation when used prod- of acceptable lots (the proposed inspection type in this
ucts are classified into a small number N of quality cate- paper), and (IT-5) no inspection, are applied to the in-
gories according to a quality classification list provided spection process in the RSC.
by a firm which remanufacturers parts/products (Guide y Used products are classified into the discrete quality
and Van Wassenhove, 2001; Guide et al., 2003; Ferguson level in IT-2, IT-3, and IT-4, so as to make application
et al., 2009; Van Wassenhove and Zikopoulos, 2010). of model analysis in this paper easier for a decision-
The motivation of this paper is to find solutions for maker of the RSC.
the following problems in the remanufacturing of used y The effect of a variety of quality in used products is
products: (i) a trade-off between inspection cost and in- considered on the probability that each part extracted
spection accuracy, (ii) the problem that remanufacture from used product is remanufacture able and the rema-
able used products are disposed, (iii) the problem that nufacturing cost depending on the quality level of each
unremanufactur able used products are conveyed to a re- used product in the model analysis of the RSC.
manufacturing process, (iv) the problem that used products y Influences of the difference in disposal costs depending
are classified into a small number N of quality categories. on the disposing stage on the maximal expected total
This paper tries to provide the optimal operation for a RSC profit of a manufacturer in the RSC are clarified.
incorporating inspection policy into remanufacturing of y This paper guarantees that no situation where used
used products to practitioners, academic researchers, and products taken as sample are inspected again occurs in
real-world policymakers who make plan for the optimal each inspection type by defining the flows of used
operation in a RSC, considering the above problems in the products and conducting the model analysis according
remanufacturing of used products in a RSC. to the clarified flows.
Concretely, the inspection process is incorporated in- y The optimal decision procedure for each inspection type
to a RSC with a single manufacturer who remanufactures is proposed to determine the optimal number of pro-
products from procured used products collected from a cured lots of used products and the optimal sampling
market. As the optimal operation for the RSC, the number inspection scheme so as to maximize the expected total
of procured lots of used products and the inspection poli- profit of a manufacturer. The optimal operation, which
cy are discussed. This paper focuses on the following five consists of the optimal number of procured lots of used
inspection types (ITs): (IT-1) 100% inspection, (IT-2) products and the optimal inspection policy, are finally
sampling inspection, (IT-3) sampling inspection with determined by comparing the maximal expected total
screening of rejected lots, (IT-4) sampling inspection with profits among five inspection types. As a result, not on-
screening of acceptable lots, and (IT-5) no inspection. IT- ly the optimal operation for the RSC but also features of
2 corresponds to a solution for the problem (i). IT-3 can each inspection type can be clarified.
resolve the demerit (I) and corresponds to a solution for The rest of this paper is organized as follows: in Sec-
the problem (ii). IT-4 can resolve the demerit (II) and tion 2, notation used in considered model is defined. In
corresponds to a solution for the problem (iii). This paper Section 3, operational flows of a RSC considered in this
determines the optimal operation for the RSC, consisting paper and model assumptions of the RSC are described.
of the optimal number of procured lots of used products In Section 4, the manufacturers expected total profit is
and the optimal inspection policy, so as to maximize the formulated for each inspection type which is incorporated
manufacturers expected total profit. into the RSC. Section 5 describes the optimal decision
The numerical analysis investigates how changes of procedures for the optimal number of procured lots of
quality distribution of used products, inspection cost of used products and the optimal sampling inspection
used products, disposal cost of used products, manufac- scheme in each inspection type. In Section 6, numerical
turing cost of new products, and the maximal number of analysis shows the optimal operation, which consists of
procured lots of used products, affect the maximal ex- the optimal number of procured lots of used products and
pected total profit of a manufacturer and the optimal op- the optimal inspection policy, in a RSC and investigates
eration in a RSC. how changes of conditions of the RSC affect the optimal
Also, from the results of numerical analysis, this pa- operation and the maximal expected total profit of a man-
per shows features of each inspection type and the usa- ufacturer in the RSC. In Section 7, conclusions, mana-
Yamaguchi and Kusukawa: Industrial Engineering & Management Systems
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 4
gerial insights and future researches for this paper are f ( x ) : the probability density function (pdf) of product
summarized. demand x
F ( x ) : the cumulative distribution function (cdf) of prod-
ucts demand x
2. NOTATION S Max : the maximal number of products which the manu-
facturer can supply in a market
y Decision Variables for a RSC v : sales price per unit of products
i : index number of inspection type (i {1, 2, , 5}) hr : the unit inventory holding cost of excess remanufac-
R : the number of procured lots of used products, referred tured products for demand x per unit time
to simply as the procured lots cM : the unit manufacturing cost of new products from
n : sample size taken from a single lot for sampling in- new parts
spection (0 n Q ) , where Q denotes the number s : shortage penalty cost per unsatisfied demand
of used products in a single lot
Qri ( R, n, Qn ) : the expected number of remanufactured
Qn : acceptance criterion for sampling inspection
products in inspection type i for procured
(0 Qn n)
lots R , sample size n , and acceptance crite-
rion Qn
y Other Variables and Functions in a RSC
P ( Qr ) : the expected profit for number of remanufactured
y : quality of used products ( 0 y 1)
products Qr , consisting of the expected sales of
g ( y ) : the probability density function (pdf) of quality y
I : the total number of quality classification products, the expected inventory holding cost of
: quality level of used products (1 I ) excess remanufactured products, the expected
G ( ) : the ratio of used products with quality level in manufacturing cost of new products, and the ex-
procured used products pected shortage penalty cost of unsatisfied prod-
E [ y | ] : the conditional expectation of quality y under uct demand
quality level TP i ( R , n, Qn ) : the expected total profit (of a single manu-
Q : the number of used products in a single lot facturer) in inspection type i for procured
RMax : the maximal number of procured lots of used lots R , sample size n , and acceptance
products criterion Qn
a : procurement cost per lot of used products R ( n, Qn ) : the optimal procured lots of used products in
i*
cc : classification cost per unit of used products inspection type i for sample size n and ac-
cd : disassembly cost per unit of used products ceptance criterion Qn
cI : inspection cost per unit of parts extracted from used ni * : the optimal sample size in inspection type i
products Qni * : the optimal acceptance criterion in inspection type i
c D1 ( n ) : the unit disposal cost of used products not taken R i * : the optimal procured lots of used products in inspec-
as sample from lots which are judged as rejected tion type i , corresponding to the optimal procured
lots in sampling inspection for sample size n , lots R i * ( ni * , Qni * ) for the optimal sample size ni *
referred to simply as the unit first disposal cost and the optimal acceptance criterion Qni * in inspec-
c D 2 : the unit disposal cost of unremanufacturable (defec- tion type i
tive) parts in inspected parts extracted from used TP : the maximal expected total profit in inspection type
i*
products including the used products taken as sam- i , corresponding to the expected total profit
ples, referred to simply as the unit second disposal TP i ( R i * , ni * , Qni * ) in inspection type i for the op-
cost timal number of procured lots R i * , the optimal
k : ratio of the minimum of cD1 ( n ) to c D 2
sample size ni * , and the optimal acceptance crite-
c D 3 : the unit disposal cost of unremanufacturable (defec-
rion Qni *
tive) parts in used products which are not inspected
Qr : the expected number of remanufactured products
i*
x : demand of a single product, referred to simply as manufactured products under optimal situation in in-
product demand spection type i
Optimal Operation for Reverse Supply Chain Incorporating Inspection Policy into Remanufacturing of Used Products
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 5
From Figure 3, as with IT-2, after the used products relation between QC and Qn , each lot is judged as ac-
are classified and are collected up at lot unit as to their ceptable lot or rejected lot. The used products ( Q n )
quality, the samples n taken from each lot are disas- not taken in the acceptable lots and the samples n are
sembled to parts and inspected. Based on the magnitude respectively treated as with IT-2. The used products
Figure 1. The flows of the operations when IT-1 is incorporated into a RSC.
Figure 2. The flows of the operations when IT-2 is incorporated into a RSC.
Figure 3. The flows of the operations when IT-3 is incorporated into a RSC.
Optimal Operation for Reverse Supply Chain Incorporating Inspection Policy into Remanufacturing of Used Products
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 7
Figure 4. The flows of the operations when IT-4 is incorporated into a RSC.
Figure 5. The flows of the operations when IT-5 is incorporated into a RSC.
( Q n ) not taken in the rejected lots are disassembled to into the products, while the defective parts are disposed in
parts and inspected. The conforming parts are conveyed the remanufacturing process at the unit cost cD 3 .
to the remanufacturing process and remanufactured into
the products, while the defective parts are disposed in the 3.3 Model Assumptions
inspection process at the unit cost cD 2 .
(1) The manufacturer doesnt know the quality y of each
y Sampling Inspection with Screening of Acceptable Lots
of used products but knows the pdf g ( y ) of quality
(IT-4): the Proposed IT
y of the whole procured used products. The higher
From Figure 4, as with IT-2 and IT-3, after the used
value of y , the higher the quality.
products are classified and collected up at lot unit as to
(2) The manufacturer doesnt know the product demand
their quality, the samples n taken from each lot are dis-
x but knows the pdf f ( x ) of demand x .
assembled to parts and inspected. Based on the magnitude
(3) It is necessary to disassemble used products to parts
relation between QC and Qn , each lot is judged as ac-
at the unit cost cd before they are inspected regard-
ceptable lot or rejected lot. Differently from IT-3, in IT-
less of IT. The inspection is conducted at the unit cost
4, the used products ( Q n ) not taken in the acceptable
c I . It enables to discriminate conforming parts from
lots are disassembled to parts and inspected. The con-
defective parts.
forming parts are conveyed to the remanufacturing
(4) Each IT doesnt conduct the destructive inspection. In
process and remanufactured into the products, while the
this case, the manufacturer can use conforming parts
defective parts are disposed in the inspection process at
which are inspected in order to remanufacture the
the unit cost cD 2 . The used products ( Q n ) not taken
products.
in the rejected lots are disposed in the inspection process
(5) As written in Section 2.2, cD1 ( n ) is the unit disposal
in a bundle at the unit cost cD1 ( n ) as with IT-2.
cost of used products not taken as samples from lots
y No Inspection (IT-5) which are judged as rejected lots in sampling inspec-
From Figure 5, after all of procured used products tion for sample size n , cD 2 is the unit disposal cost
are disassembled to parts without classification as to qual- of defective parts in inspected parts extracted from
ity and inspection, they are conveyed to the remanufactur- used products including the used products taken as
ing process. The conforming parts are remanufactured samples, and cD 3 is the unit disposal cost of defec-
Yamaguchi and Kusukawa: Industrial Engineering & Management Systems
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 8
{ }
In this case, the expected number of the conform-
P ( Qr ) = v
S Max
Q p {N ,
QC = 0
C QC , r ( )} cM
S Max
Qr
( x Qr ) f ( x ) dx
Optimal Operation for Reverse Supply Chain Incorporating Inspection Policy into Remanufacturing of Used Products
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 9
( x SMax ) f ( x ) dx
I
s cd ( Q n ) Bn ( Qn , n, 1, ) RG ( )
S Max
{ F ( x ) dx} h F ( x ) dx
=1
S Max Qr
= v S Max I
cD 3 ( Q n ){1 r ( )} Bn ( Qn , n, 1, ) RG ( )
0 r 0
c {S Q F ( S ) F ( x ) dx}
S Max =1
M Max r Max I
cr (
Qr
)( Q n ) r ( ) Bn ( Qn , n, 1, ) RG ( )
s {E [ x ] S + F ( x ) dx}
S Max
Max (6) =1
+ P {Qr2 ( R, n, Qn )} ,
0
(9)
(See Appendix B).
I
4.1 Case of 100% Inspection (IT-1) Qr2 ( R, n, Qn ) = nr ( ) RG ( )
=1
I
From Section 3.2, both of n and Qn are 0 in IT-1. + ( Q n ) r ( ) Bn ( Qn , n, 1, ) RG ( ) (10)
Therefore, the expected total profit TP1 ( R, 0, 0 ) and the =1
spectively formulated as ( cd + cI ) ( Q n ) Bn ( 0, Qn 1, 1, ) RG ( )
=1
I
TP 2 ( R, n, Qn ) = aR ccQR (cd + cI )nR cD 2 ( Q n ){1 r ( )} Bn ( 0, Qn 1, 1, ) RG ( )
I =1
cD 2 n {1 r ( )} RG ( ) I
=1
cr ( )( Q n ) r ( ) Bn ( 0, Qn 1, 1, ) RG ( )
=1
I
cr ( ) nr ( ) RG ( )
I
cd ( Q n ) Bn ( Qn , n, 1, ) RG ( )
=1
=1
I I
cD1 (n) (Q n) Bn (0, Qn 1, 1, ) RG ( ) cD 3 ( Q n ){1 r ( )} Bn ( Qn , n, 1, ) RG ( )
=1 =1
Yamaguchi and Kusukawa: Industrial Engineering & Management Systems
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 10
I
The terms from first to sixth in Eq. (13) are same as
cr ( )( Q n ) r ( ) Bn ( Qn , n, 1, ) RG ( )
=1
those in Eq. (9). In Eq. (13), the seventh term is the ex-
, (11) pected disassembly cost and the expected inspection cost
+ P {Qr3 ( R, n, Qn )} of parts extracted from the used products not taken as
sample in acceptable lots. The eighth term is the expected
I
second disposal cost of defective parts not taken as sam-
Qr3 ( R, n, Qn ) = Qr ( ) RG ( ) (12)
=1
ple in the acceptable lots. The ninth term is the expected
remanufacturing cost of products from conforming parts
The terms from first to fifth in Eq. (11) are same as not taken as sample in the rejected lots. The tenth term is
those in Eq. (9). The sixth term is the expected disassem- calculated from Eq. (6) in IT-4.
bly cost and the expected inspection cost of parts not tak-
en as sample in the rejected lots. The seventh term is the 4.5 Case of No Inspection (IT-5)
expected second disposal cost of defective parts not taken
as sample in the rejected lots. The eighth term is the ex- From Section 3.2, both of n and Qn are 0 in IT-5.
pected remanufacturing cost of products from conforming Therefore, the expected total profit TP 5 ( R, n, Qn ) and
parts not taken as sample in the rejected lots. The ninth the expected number of remanufactured products Qr5
term is the expected disassembly cost of used products ( R, n, Qn ) in IT-5 for R are respectively formulated as
not taken as sample in the acceptable lots. The tenth term
is the expected third disposal cost of defective parts not I
TP 5 ( R, 0, 0 ) = aR cd QR cD 3 Q {1 r ( )} RG ( )
taken as sample in the acceptable lots. The eleventh term =1
is the remanufacturing cost of products from conforming I
parts not taken as sample in the acceptable lots. The cr ( ) Qr ( ) RG ( ) + P {Qr5 ( R, 0, 0 )} (15)
=1
twelfth term is calculated from Eq. (6) in IT-3. Note that
Qr3 ( R , n, Qn ) doesnt depend on n and Qn because any
I
conforming part isnt disposed in IT-3. Qr5 ( R, 0, 0 ) = Qr ( ) RG ( ) . (16)
=1
parts extracted from used products such as the uniform the inspection cost, the first disposal cost of used products,
distribution-type shape, the normal distribution-type the second disposal cost and the third disposal cost of
shape, the exponential distribution-type shape, the left- parts, the remanufacturing cost of products, and the in-
biased distribution shape, the right-biased distribution ventory holding cost of remanufactured products.
shape, by using the following probability density function The maximum expected total profit is high in the
of quality, g ( y ) with parameters ( ma , mb ) : order of IT-2, IT-4, IT-1, IT-3, and IT-5. Accordingly, the
optimal operation in a RSC is finally obtained as
( ma + mb ) ( R* , i* , n* , Qn* ) = (161, 2, 12, 5) . In addition, the expected
g ( y) = y ma (1 y ) ( 0 y 1)
mb
(23)
( ma ) ( mb ) number of remanufactured products in IT-2, Qr2 * is the
largest among five ITs. Therefore, IT-2 enables to pro-
where ( ) denotes the gamma function. mote the recycling activity of used products and to im-
The first disposal cost cD1 ( n ) for sample size n is set prove the maximum expected total profit. The maximum
as cD1 ( n ) = (1 k ) cD 2 n / ( Q 1) + kcD 2 , where 0 n Q 1 , expected total profit in sampling inspection with screen-
satisfying the conditions: cD1 ( n ) cD 2 and cD1 ( Q 1) ing of acceptable lots IT-4, which is the proposed IT in
= cD 2 in Section 3.3 (5). Here, k is a constant within
this paper, is higher than that in IT-1. The expected num-
0 k 1 , and k represents a ratio of the minimum of
ber of remanufactured products in IT-4 is about 99.9% of
cD1 ( n ) to cD 2 according to cD1 ( 0 ) = kcD 2 .
that in IT-1.
The conforming probability and the remanufacturing
6.2.1 Effect of Quality Distribution of Used Products
cost of a part with quality level are respectively set as
on Optimal Operation in a RSC
r ( ) = 0.5 0.4cos ( E [ y | ] ) and
By changing parameters ( ma , mb ) in pdf g ( y ) of
cr ( ) = 20 E [ y | ] + 40 , satisfying the properties of func-
the beta distribution in Eq. (23), the following six cases of
tions for quality level in Section 2.3 (9). Here, E [ y | ]
the quality distribution of used products are provided as:
denotes the conditional expectation of quality level y
Case 1: ( ma , mb ) = (1, 1) : a situation where the quality
under quality level . E [ y | ] is calculated as
level of used products is distributed uniformly,
Case 2: ( ma , mb ) = ( 2, 2 ) : a situation where there are
yg ( y )
E[ y | ] = ( 1) / I
/I
dy (24) many used products with the middle quality lev-
G( ) el,
Case 3: ( ma , mb ) = ( 2, 3) : a situation where there are
The following numerical example are used as system many used products with the relatively low
parameters in a RSC: Q = 100, I = 20, a = 1000, cc = 3, quality level,
cd = 5, cI = 15, k = 1 4, cD 2 = 30, cD 3 = 65, v = 180, hr = 10, Case 4: ( ma , mb ) = ( 3, 5) : a situation where there are
cM = 160, s = 10, RMax = 200, SMax = 20000, = 7000, many used products with the low quality level,
= 500, ma = 2, mb = 2. Case 5: ( ma , mb ) = ( 3, 2 ) : a situation where there are
many used products with the relatively high
6.2 Results of the Numerical Analysis quality level,
Case 6: ( ma , mb ) = ( 5, 3) : a situation where there are
Table 1 shows the results of the optimal operation in many used products with the high quality level.
each IT obtained from the numerical example.
Here, cri * shown in Table 1 indicates the average of The probability density function g ( y ) of quality
the expected total cost per unit of remanufactured prod-
ucts under the optimal situation in IT-i. The expected total
cost related to remanufactured products is calculated from
the sum of the following expected values of the procure-
ment cost, the classification cost, the disassembly cost,
y in each case is shown in Figure 6. products from conforming parts and manufactures new
Table 2 shows the effect of the quality distribution products from new parts to satisfy the product demand.
on the optimal operation in each IT. y Under Case 4 of the quality distribution, the manufac-
From Table 2, the following results can be seen: turer doesnt remanufacture products from used prod-
y Under case 1 of the quality distribution, the maximum ucts in all ITs.
expected total profit of a manufacturer is high in the or- y Under Case 5 of the quality distribution, IT-2 is the op-
der of IT-2, IT-4, IT-1, IT-3, and IT-5 as with the results timal IT. In addition, the optimal number of procured
obtained from numerical examples in Table 1 where the lots R i * in all ITs is smaller than that under Case 2
quality distribution is Case 2. provided as the numerical example. This is because
y Under Case 3 of the quality distribution, inspection type mean of the conforming probability of used products is
1 is the optimal IT. The manufacturers expected total so higher that the required number of used products to
profit in proposed inspection type, IT-4, is the second satisfy the product demand is smaller.
highest. Also, in IT-5, the optimal procured lots R 5* = 0 . y Under Case 6 of the quality distribution, R i * is further
In this case, the manufacturer doesnt remanufacture smaller than that under Case 5 of that. Then, IT-5 is the
optimal IT.
Table 2. Effect of quality distribution of used products on y IT-4 is the worst rank of the manufacturers maximal
optimal operation in each IT expected profit under Cases 5 and 6 of the quality dis-
tribution.
g ( y) i Ri * ni * Qni * TPi * Rank Qri * Therefore, sampling inspection with screening ac-
1 133 0 0 415,393 3rd 6,650 ceptable lots, which is IT-4 (the proposed inspection type),
2 170 8 4 494,482 1 st
6,762
is effective under either situations where the quality level
of used products is uniformly distributed or there are
Case 1 3 133 10 6 410,739 4th 6,650
many used products with the middle or low quality level.
4 160 7 3 456,541 2nd 6,713
5 132 0 0 382,158 5th 6,600 6.2.2 Effect of Unit Inspection Cost on Optimal Op-
1 133 0 0 407,777 3rd 6,650 eration in a RSC
Table 3 shows effect of cI on optimal operation in
2 161 12 5 425,340 1st 6,689
each IT. Note that the optimal operation in IT-5 is con-
Case 2 3 133 13 8 392,154 4th 6,650 stant for cI due to no inspected parts.
4 153 11 4 409,179 2nd 6,644 From Table 3, the following results can be seen.
5 132 0 0 374,600 5th 6,600 y The higher the unit inspection cost cI is, the smaller
1 156 0 0 188,802 1st 6,161 the optimal sample size ni * and the optimal acceptance
2 188 15 5 161,785 3rd 6,001
Case 3 3 149 19 12 153,506 4th 5,884 Table 3. Effect of the unit inspection cost cI on optimal
4 181 11 3 184,313 2nd 6,154 operation in each IT
5 0 0 0 140,000 5th 0 cI i Ri * ni * Qni * TPi * Rank Qri *
1 0 0 0 140,000 1st 0 1 135 0 0 471,828 1st 6,750
st
2 0 0 0 140,000 1 0 2 163 14 6 435,627 4th 6,716
Case 4 3 0 0 0 140,000 1st 0 10 3 134 21 16 440,975 3rd 6,700
4 0 0 0 140,000 1st 0 4 151 12 4 466,007 2nd 6,716
5 0 0 0 140,000 1st 0 5 132 0 0 374,600 5th 6,600
1 113 0 0 557,250 4th 6,837 1 133 0 0 407,777 3rd 6,650
2 129 11 5 575,675 1st 6,876 2 161 12 5 425,340 1st 6,689
Case 5 3 113 10 6 557,460 3rd 6,837 15 3 133 13 8 392,154 4th 6,650
4 124 10 4 541,629 5th 6,841 4 153 11 4 409,179 2nd 6,644
5 113 0 0 570,560 2nd 6,837 5 132 0 0 374,600 5th 6,600
1 108 0 0 591,118 4th 6,878 1 131 0 0 341,676 5th 6,550
2 119 11 5 597,140 2nd 6,864 2 165 9 4 417,582 1st 6,654
Case 6 3 108 9 5 592,100 3rd 6,878 20 3 132 8 4 355,600 3rd 6,600
4 112 14 5 564,932 5th 6,835 4 156 10 4 353,430 4th 6,565
5 108 0 0 615,850 1st 6,878 5 132 0 0 374,600 2nd 6,600
Yamaguchi and Kusukawa: Industrial Engineering & Management Systems
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 14
criterion Qni * tend to be in IT-2, IT-3, and IT-4. introducing IT-2, IT-3, and IT-4.,
y As cI increases, the optimal procured lots R i * in IT-3
and IT-5 decreases and that in IT-4 increases. Here, in 6.2.3 Effect of Unit First Disposal Cost on Optimal
IT-4, a manufacturer incurs the inspection cost of parts Operation in a RSC
( Q n ) not take as samples in the acceptable lots. Thus, By varying the constant k in the unit first disposal
in IT-4 with a strict inspection, the more conforming cost cD1 ( n ) , a sensitivity analysis with respect to cD1 ( n )
parts are disposed, and the more used products are ne- is conducted.
cessary to remanufacture products. Table 4 shows effect of k on optimal operation in
y IT-1 and IT-4 are largely affected with the change of the each IT. Note that the optimal operation in IT-1, IT-3, and
unit inspection cost cI than IT-2 since they need to IT-5, is unaffected with k according to Section 3.2.
conduct an inspection for more used products. From Table 4, the following results can be seen.
y The rank order in IT-4 falls from the second to the forth y In IT-2 and IT-4, as k becomes larger, the maximal
while cI changes from 10 to 20. The rank order in IT- expected total profit decreases.
1 falls from the first to the worst then. It is indicated y While k changes from 0 to 1, the rank order in IT-4 falls
that IT-4 is less affected with cI than IT-1 in Case 2. from the second to the forth. The rank order in IT-2 falls
y When cI is high ( cI = 20 ) , Qr1* < Qri * ( i = 2,3,4 ) is from the first to the worst then. It is indicated that IT-4
satisfied. It is verified that the effect of increment in cI is less affected with k than IT-2 in Case 2.
on the maximal expected total profit and the expected y Differently from IT-4, IT-2 has the potential to convey
number of remanufactured products can be relieved by defective parts to the remanufacturing process. It indi-
cates that IT-2 poses a risk that a manufacturer incurs
Table 4. Effect of the constant k in the first disposal cost
cD1 ( n ) of used products on optimal operation in Table 5. Effect of the second disposal cost cD 2 of defec-
each IT tive parts on optimal operation in each IT
k i Ri * ni * Qni * cD1 ( n* ) TPi * Rank Qri * cD 2 i Ri * ni * Qni * TPi * Rank Qri *
the third disposal cost, which is the highest of three types constant for cD 3 since cD 3 is the unit disposal cost of defec-
of disposal cost, when defective parts are detected in the tive parts conveyed to the remanufacturing process.
remanufacturing process. When the first disposal cost is From Table 6, the following results can be seen.
higher due to the increases of k , the optimal sample size y When cD 3 is high ( cD 3 = 75, 85) , IT-4, which is the
n i * becomes larger to avoid the increase of disposal cost. proposed IT, is the optimal IT. It is verified that it is
However, as a result, the inspection cost increases as the profitable to incorporate IT-4 into a RSC under a situa-
optimal sample size ni * becomes larger. tion where the disadvantage is significant when defec-
tive parts are conveyed to the remanufacturing process.
6.2.4 Effect of Unit Second Disposal Cost on Optimal Therefore, it is worth adding IT-4 to choices of IT.
Operation in a RSC y When cD 3 = 85 , the optimal sample size in IT-2 is
Table 5 shows effect of the unit second disposal cost n 2* = 100 ( = Q ) . It indicates that all of parts extracted
cD 2 on optimal operation in each IT. Note that the optim- from procured used products are inspected and only
al operation in IT-5 is unaffected with cD 2 according to conforming parts are conveyed to the remanufacturing
Section 3.2. process. The maximum expected total profit in IT-2
From Table 5, the following results can be seen. with classification cost is lower than that in IT-1 with-
y In IT-1, IT-2, IT-3, and IT-4, as cD 2 becomes larger, the out it.
maximal expected total profit decreases. y When cD 3 is low ( cD 3 = 50 ) , the optimal sample size
y As with the sensitivity analysis with respect to cI , the in IT-3 is n3* = 0 . It indicates that all of parts extracted
rank order in IT-4 falls from the second to the third from procured used products are conveyed to the rema-
while cD 2 changes from 20 to 40. However, IT-4 is nufacturing process without inspection. The maximum
less affected with cD 2 than IT-1 in Case 2. expected total profit in IT-3 with classification cost is
lower than that in IT-5 without it.
6.2.5 Effect of Unit Third Disposal Cost on Optimal
Operation in a RSC 6.2.6 Effect of Manufacturing Cost of New Products
Table 6 shows effect of cD 3 on optimal operation in on Optimal Operation in a RSC
each IT. Note that the optimal operation in IT-1 and IT-4 is Table 7 shows the effect of cM on the optimal op-
eration in each IT.
Table 6. Effect of the third disposal cost cD 3 of defective From Table 7, the following results can be seen:
parts on optimal operation in each IT y The rank order of the IT and the optimal inspection
scheme are unaffected with the change of the unit man-
cD 3 i Ri * ni * Qni * TPi * Rank Qri * ufacturing cost cM .
1 133 0 0 407,777 5th 6,650 y When cM is low ( cM = 115, 120 ) , R i * and Qri * become 0.
2 157 8 3 479,675 1st 6,739 This result indicates that the new production from new
parts to new products tends to be promoted as cM de-
50 3 134 0 0 434,489 3rd 6,700
creases.
4 153 11 4 409,179 4th 6,644 y The higher cM is, the larger the optimal procured lots
5 135 0 0 474,818 2nd 6,750 R i * and the expected remanufactured products Qri * are.
1 133 0 0 407,777 3rd 6,650 This result indicates that the remanufacturing activity
2 161 12 5 425,340 1st 6,689 from used products to products tends to be promoted as
cM becomes higher.
65 3 133 13 8 392,154 4th 6,650
y When IT-3 or IT-5 is incorporated into a RSC, the re-
4 153 11 4 409,179 2nd 6,644 manufacturing activity from used products tends to be
5 132 0 0 374,600 5th 6,650 stopped in the earlier stage than other ITs as cM be-
1 133 0 0 407,777 2nd 6,650 comes lower. This means that the total cost related to
2 167 13 6 393,745 3rd 6,645 the unit of remanufactured products is higher than that
75 3 132 17 12 382,137 4th 6,600
in the other inspection types. This can be confirmed
from the result in Table 1 that cri * in IT-3 and IT-5 are
4 153 11 4 409,179 1st 6,644
higher than those in other ITs.
5 130 0 0 309,025 5th 6,500
1 133 0 0 407,777 2nd 6,650 6.2.7 Effect of Maximal Number of Procured Lots
2 132 100 0 368,000 4th 6,600 on Optimal Operation in a RSC
In this section, by changing the maximum number of
85 3 132 22 17 376,413 3rd 6,600
procured lots of used products RMax , situations where the
4 153 11 4 409,179 1st 6,644 manufacturer cant procure so many used products are
5 127 0 0 244,673 5th 6,350 discussed.
Yamaguchi and Kusukawa: Industrial Engineering & Management Systems
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 16
Table 7. Effect of the unit manufacturing cost cM of Table 8. Effect of the maximal number of procured lots of
new products on optimal operation in each IT used products RMax on optimal operation in each
IT
cM i Ri * ni * Qni * TPi * Rank Qri *
RMax i Ri * ni * Qni * TPi * Rank Qri *
1 0 0 0 455,000 1st 0
2 0 0 0 455,000 1st 0 1 130 0 0 406,525 1st 6,500
115 3 0 0 0 455,000 1st 0 2 130 16 5 391,915 2nd 5,893
4 0 0 0 455,000 1st 0 130 3 130 13 8 391,255 3rd 6,500
5 0 0 0 455,000 1st 0 4 130 15 4 388,814 4th 6,052
1 119 0 0 432,037 3 rd
5,950 5 130 0 0 374,025 5th 6,500
2 147 12 5 447,901 1st 6,107 1 133 0 0 407,777 2nd 6,650
120 3 0 0 0 420,000 4th 0 2 140 15 5 408,843 1st 6,250
nd
4 137 11 4 433,300 2 5,949 140 3 133 13 8 392,154 4th 6,650
5 0 0 0 420,000 4th 0 4 140 14 4 402,412 3rd 6,440
1 124 0 0 427,412 3rd 6,200 5 132 0 0 374,600 5th 6,600
2 151 12 5 443,815 1st 6,273 1 133 0 0 407,777 3rd 6,650
125 3 122 13 8 412,976 4th 6,100 2 161 12 5 425,340 1st 6,689
4 143 11 4 428,724 2nd 6,210 200 3 133 13 8 392,154 4th 6,650
5 118 0 0 397,023 5th 5,900 4 153 11 4 409,179 2nd 6,644
1 132 0 0 409,900 3rd 6,600
5 132 0 0 374,600 5th 6,600
2 160 12 5 427,389 1st 6,647
155 3 132 13 8 394,393 4th 6,600
IT-4, is effective when under the situations where the
4 153 11 4 411,308 2nd 6,644
quality level of used products is uniformly distributed or
5 131 0 0 376,927 5th 6,550
rd
there are many used products with the middle or low
1 133 0 0 407,777 3 6,650 quality level. Moreover, in Sections 6.2.2-6.2.4, IT-4 is
2 161 12 5 425,340 1st 6,689 detected to be less affected with cI , k , and cD 2 . How-
160 3 133 13 8 392,154 4th 6,650 ever, these sensitivity analyses are performed only under
4 153 11 4 409,179 2nd 6,644 the situation where there are many used products with the
5 132 0 0 374,600 5th 6,600 middle quality level, Case 2. This section performs the
1 136 0 0 400,343 3rd 6,800 sensitivity analyses with respect to cI , k , and cD 2 , in
2 164 12 5 418,222 1st 6,813 Case 3 as a situation where the quality of many used
180 3 135 13 8 384,481 4th 6,750 products is worse.
Table 9-Table 11 show the results of the sensitivity
4 156 11 4 401,822 2nd 6,775
analyses with respect to cI , k , and cD 2 , in Case 3, re-
5 135 0 0 366,590 5th 6,750
spectively.
From Table 9, it can be seen that IT-4 is less affected
Table 8 shows the effect of RMax on the optimal with cI than IT-1 in Case 3 as with Case 2. In addition,
operation in each IT. when cI = 17 , IT-4 is the optimal IT. However, when
From Table 6, it can be seen that as RMax decreases, cI = 20 , the manufacturer doesnt procure used products
the maximal expected total profit in IT-2 and IT-4 be-
except in IT-2.
comes lower significantly. This is because IT-2 and IT-4
When k = 1 2, 3 4, 1, in Table 10, or cD 2 = 20, 25,
have the following property: many used products are pro-
cured and more used products with high quality level at in Table 11, all of parts extracted from procured used
lot unit are conveyed to the remanufacturing process. This products are inspected in IT-2 (i.e. n 2* = 100 ). Under a
can bring more profits to a manufacturer than the other situation where low quality used products are many, it is
ITs. Therefore, it is clarified that IT-2 and IT-4 are af- more difficult for IT-2, IT-3, and IT-5, to function because
fected more greatly with RMax than other inspection types. defective parts are conveyed to the remanufacturing
process in these ITs. Also, IT-4 is no absolute exception,
6.2.8 Sensitivity Analyses in Case 3 and IT-1 tends to be more effective and less affected with
In Section 6.2.1, it is indicated that the proposed IT, k and cD 2 than IT-4 under the situation.
Optimal Operation for Reverse Supply Chain Incorporating Inspection Policy into Remanufacturing of Used Products
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 17
Table 9. Sensitivity analysis with respect to cI in Case 3 Table 11. Sensitivity analysis with respect to cD 2 in Case 3
i* i*
cI i Ri *
i*
n Qn TP i * Rank Qr cD 2 i Ri * ni * Qni * TP i * Rank Qri *
1 163 0 0 268,659 1st 6,437 1 163 0 0 285,788 1st 6,437
2 159 100 0 220,366 4th 6,279 2 161 100 0 237,168 4th 6,358
10 3 159 30 23 222,363 3rd 6,279 20 3 161 25 18 242,426 3rd 6,358
4 188 11 3 251,624 2nd 6,392 4 191 14 4 273,723 2nd 6,430
5 0 0 0 140,000 5th 0
5 0 0 0 140,000 5th 0
1 156 0 0 188,802 1st 6,161
1 161 0 0 236,758 1st 6,358
2 188 15 5 161,785 3rd 6,001
15 3 149 19 12 153,506 4th 5,884 2 156 100 0 189,199 4th 6,161
4 181 11 3 184,313 2nd 6,154 25 3 157 22 15 196,629 3rd 6,200
5 0 0 0 140,000 5th 0 4 188 14 4 228,441 2nd 6,329
1 151 0 0 158,031 2nd 5,963 5 0 0 0 140,000 5th 0
2 192 14 5 156,304 3rd 5,936 1 156 0 0 188,802 1st 6,161
17 3 0 0 0 140,000 4th 0 2 188 15 5 161,785 3rd 6,001
4 181 10 3 158,660 1st 5,969 30 3 149 19 12 153,506 4th 5,884
5 0 0 0 140,000 4th 0 4 181 11 3 184,313 2nd 6,154
1 143 0 0 143,247 3rd 5,647 5 0 0 0 140,000 5th 0
2 186 12 4 153,757 1st 5,909 1 142 0 0 142,884 1st 5,608
18 3 0 0 0 140,000 4th 0 2 177 12 4 142,757 2nd 5,623
4 175 10 3 146,306 2nd 5,771 35 3 0 0 0 140,000 4th 0
5 0 0 0 140,000 4th 0 4 169 10 3 142,215 3rd 5,574
1 0 0 0 140,000 2nd 0 5 0 0 0 140,000 4th 0
2 191 11 4 149,487 1st 5,830 1 0 0 0 140,000 1st 0
20 3 0 0 0 140,000 2nd 0
2 0 0 0 140,000 1st 0
4 0 0 0 140,000 2nd 0
40 3 0 0 0 140,000 1st 0
5 0 0 0 140,000 2nd 0
4 0 0 0 140,000 1st 0
5 0 0 0 140,000 1st 0
Table 10. Sensitivity analysis with respect to k in Case 3
k i Ri *
i*
ni * Qn cD1 ( n* ) TP i * Rank Qri *
1 156 0 0 * 188,802 3rd 6,161
2 200 10 4 3.03 214,448 2nd 5,798 7. CONCLUSION
0 3 149 19 12 * 153,506 4th 5,884
4 200 8 3 2.42 229,117 1st 6,012
This paper incorporated the following five inspec-
5 0 0 0 * 140,000 5th 0 tion type (IT)s into a RSC which consisted of the process
1 156 0 0 * 188,802 1st 6,161 flows from procurement of used products collected from
2 188 15 5 10.91 161,785 3rd 6,001 a market, through remanufacturing products from the
1 used products, to sales of the products in a market: (IT-1)
3 149 19 12 * 153,506 4th 5,884
4 100% inspection, (IT-2) sampling inspection, (IT-3) sam-
4 181 11 3 10 184,313 2nd 6,154
5 0 0 0 * 140,000 5th 0 pling inspection with screening of rejected lots, (IT-4)
1 156 0 0 * 188,802 1st 6,161 sampling inspection with screening of acceptable lots(the
2 143 100 0 30.15 143,247 4th 5,647 proposed IT), and (IT-5) no inspection.
1 This paper formulated the manufacturers expected
3 149 19 12 * 153,506 3rd 5,884
2 total profit in each IT and derived a procedure to deter-
4 164 19 4 17.88 158,290 2nd 5,968
5 0 0 0 * 140,000 5th 0 mine the optimal operation for a procurement of used
1 156 0 0 * 188,802 1st 6,161 products and an inspection policy, which maximized the
2 143 100 0 30.08 143,247 4th 5,647 manufacturers expected total profit, by theoretical analy-
3
3 149 19 12 * 153,506 2nd 5,884 sis and numerical search.
4
4 150 28 4 24.62 146,339 3rd 5,753 In addition, this paper considered the following top-
5 0 0 0 * 140,000 5th 0 ics:
1 156 0 0 * 188,802 1st 6,161 y Classification of used products into the discrete quality
2 143 100 0 30 143,247 4th 5,647 level,
1 3 149 19 12 * 153,506 2nd 5,884 y The probability that each part extracted from used prod-
4 144 61 6 30 143,429 3rd 5,664 ucts is remanufacturable and the remanufacturing cost
5 0 0 0 * 140,000 5th 0 depending on the quality level of each part,
Yamaguchi and Kusukawa: Industrial Engineering & Management Systems
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 18
y The difference in disposal costs depending on the dis- y The optimal decision for inspection policy for lots with
posing stage in a RSC, each quality level in used products,
y No situation where used products taken as sample are y The optimal decision for optimal inspection policy con-
inspected again. sidering not only the expectation of manufacturers
The numerical analysis investigated how changes of profit, but also a variance of it,
quality distribution of used products, inspection cost of y The optimal decision for optimal inspection policy
used products, dispose cost of used products, manufactur- when a RSC consists of a third-party-collection trader
ing cost of new products, and the maximal number of who collects used products from markets/customers and
procured lots of used products, affected the maximal ex- sorts them and manufacturer who inspects them by five
pected total profit of a manufacturer and the optimal op- ITs discussed in this paper,
eration in a RSC. Also, from the results of numerical y Consideration for destructive inspection and deteriora-
analysis, this paper showed the usability to add IT-4, tion in quality.
which was the proposed inspection in this study, to choic-
es of inspection type for the RSC. Furthermore, this paper
determined the conclusive optimal operation for a RSC ACKNOWLEDGEMENT
by comparing the maximal expected total profits among
five ITs. Thus, the numerical analysis clarified the fea- This research has been supported by the Grant-in-
tures of each IT. Aid for Scientific Research C No. 25350451 from the
Results of numerical analysis in this paper provided Japan Society for the Promotion of Science.
the following managerial insights.
y The appropriate IT should be chosen by a manufacturer
who inspects and remanufactures the procured used REFERENCES
products as to conditions of the RSC.
y Adding the proposed IT in this paper to choices of IT Aras, N., Boyaci, T., and Verter, V. (2004), The effect of
for a RSC could bring improvement in the manufactur- categorizing returned products in remanufacturing,
ers expected total profit. IIE Transactions, 36(4), 319-331.
y Especially in a situation where there are many used Behret, H. and Korugan, A. (2009), Performance analysis
products with middle quality level, the manufacturers of a hybrid system under quality impact of returns,
expected total profit in IT-4, the proposed IT, tended to Computers & Industrial Engineering, 56(2), 507-
be less affected with changes of conditions of the RSC 520.
including the unit inspection cost and the unit disposal Fleischmann, M., Bloemhof-Ruwaard, J. M., Dekker, R.,
cost. Van der Laan, E., Van Nunen, J. A., and Van Was-
However, this paper did not discuss the optimal in- senhove, L. N. (1997), Quantitative models for re-
spection policy for lots with each quality level. verse logistics: A review, European Journal of Op-
Also, the optimal decision for optimal inspection erational Research, 103, 1-17.
policy was determined so as to maximize the expected Ferguson, M., Guide, V. D., Koca, E., and Souza, G. C.
profit of a manufacturer, considering only the expectation (2009), The value of quality grading in remanufac-
of manufacturers profit. When a third-party-collection turing, Production and Operations Management, 18,
trader which collects used products from markets/cus- 300-314.
tomers and sorts them is incorporated into the RSC as a Guide, V. D. R. J. R. (2000), Production planning and
procurement source, it is necessary for a manufacturer to control for remanufacturing: Industry practice and
determine newly not only the optimal procurement quan- research needs, Journal of Operation Management, 18,
tity, but also the optimal inspection policy in a RSC, con- 467-483.
sidering quality of used products. Guide, V. D. R., Teunter, R. H., and Van Wassenhove, L.
Moreover, to apply this model to different type of N. (2003), Matching demand and supply to maxim-
products from consumer electronics and electronic com- ize profits from remanufacturing, Manufacturing &
ponents, occasionally destructive inspections should be Service Operations Management, 5, 303-316.
adopted. By considering the destructive inspection and Guide, V. D. R. and Van Wassenhove, L. N. (2001),
deterioration in quality, this model may be applied to food Managing product returns for remanufacturing,
products, because the primary materials for food products Production and Operations Management, 10(2), 142-
also have the high uncertainty in quality as with used 155.
products in RSC. Inderfurth, K. (2005), Impact of uncertainties on recovery
From the above, as future researches, it will be ne- behavior in a remanufacturing environment: A nu-
cessary to incorporate the following topics into the optim- merical analysis, International Journal of Physical
al operation policy for a RSC model in this paper: Distribution & Logistics Management, 35(5), 318-
Optimal Operation for Reverse Supply Chain Incorporating Inspection Policy into Remanufacturing of Used Products
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 19
APPENDIX A s { ( x S
0 Max
SMax
) f ( x ) dx 0 ( x SMax ) f ( x ) dx }
= v {S F ( S }
The elicitation process of Eq. (3) is shown as follows. S Max
Qr }
F ( x ) dx
N ( N 1)!
=
N 1
= N
( QC 1)!{( N 1) ( QC 1)}! QC 1 {
s E [ x ] S Max xF ( x ) 0
S Max
c {S Q F ( S ) F ( x ) dx}
SMax
N
Q p { N , Q , r ( )}
M Max r Max Qr
s {E [ x ] S + F ( x ) dx}
C C
QC = 0 S Max
Max
N N 0
Q {r ( )} {1 r ( )}
QC N QC
= C
QC = 0 QC
Therefore, Eq. (6) has been proved.
N
= 0 {r ( )} {1 r ( )}
0 N 0
0
N
N 1 APPENDIX C
+ N {r ( )} {1 r ( )}
QC N QC
QC =1 QC 1
The elicitation process of Eq. (20) which indicates
N N 1 that TP i ( R, n, Qn ) is a concave function with respect to
{1 r ( )}(
N 1) ( QC 1)
) {r ( )}
QC 1
= Nr (
QC =1 QC 1 R under n and Qn , is shown as follows.
N 1 N 1 ( N 1) QC Due to
= Nr ( ) {r ( )} {1 r ( )}
QC
QC = 0 QC
. Qri ( R , n , Qn ) ,
) {r ( )} + {1 r ( )}
N 1
= Nr ( = Nr ( ) 1N 1 = Nr ( ) F ( x ) dx = F {Qri ( R, n, Qn )} Qri ( R, n, Qn )
R 0 R
i
Therefore, Eq. (3) has been proved. Qr ( R, n, Qn ) = u i ( n, Qn ) ,
R
2 i
Qr ( R, n, Qn ) = 0 ,
R 2
APPENDIX B
The elicitation process of Eq. (6) is shown as follows. the first- and second-order partial derivatives of TP i
( R, n, Qn ) with respect to R are respectively derived as
P ( Qr )
{ }
=v
S Max
xf ( x ) dx +
S Max f ( x ) dx TP i ( R, n, Qn )
0 SMax R
i
= t i ( n, Qn ) hr F {Qri ( R, n, Qn )} Qr ( R, n, Qn )
Qr
hr
0
( Qr x ) f ( x ) dx R
SMax
cM
Qr
( x Qr ) f ( x ) dx
cM F ( S Max ) Qri ( R, n, Qn )
R
s ( x S Max ) f ( x ) dx
SMax
+ F {Qri ( R, n, Qn )} Qri ( R, n, Qn )
{
= v xF ( x ) 0
S Max
0
S Max
F ( x ) dx
R
= t i ( n, Qn ) ( hr + cM ) F {Qri ( R, n, Qn )}
i
Qr ( R, n, Qn )
+ S Max f ( x ) dx
0 0
S Max
S Max f ( x ) dx } i
R
+ cM F ( S Max ) Qr ( R, n, Qn )
{
hr Qr F ( x ) 0 xF ( x ) 0 + F ( x ) dx
Qr
0
Qr Qr
} R
= t i ( n, Qn ) ( hr + cM ) F { Ru i ( n, Qn )} u i ( n, Qn )
c { xF ( x ) }
S Max
F ( x ) dx Qr F ( x ) Q
S Max S Max
M Qr Qr r +cM F ( S Max ) u i ( n, Qn ) ,
Optimal Operation for Reverse Supply Chain Incorporating Inspection Policy into Remanufacturing of Used Products
Vol 16, No 1, March 2017, pp.1-21, 2017 KIIE 21
2
TPi ( R, n, Qn ) ( hr + cM ) F {Ri * ( n, Qn ) u i ( n, Qn )}
R 2
= t i ( n, Qn ) + cM F ( S Max ) u i ( n, Qn )
.
= ( hr + cM ) f {Ru i ( n, Qn )}{u i ( n, Qn )} < 0
2
F { R i * ( n, Qn ) u i ( n, Qn )}
Therefore, Eq. (20), indicating that TP i ( R, n, Qn ) is t i ( n, Qn ) c F ( S Max )
a concave function with respect to R under n and Qn , = + M ( hr + cM > 0 )
( hr + cM ) u i ( n, Qn ) hr + cM
is proved.
R i * ( n, Qn )
1 t i ( n, Qn ) c F ( S Max )
= i F 1 + M
APPENDIX D u ( n, Qn ) ( hr + cM ) u ( n, Qn )
i
hr + cM