Applied Mathematics and Computation 256 (2015) 742753

Contents lists available at ScienceDirect

Applied Mathematics and Computation

journal homepage: www.elsevier.com/locate/amc

Remanufacturing production planning with compensation

function approximation method
Haijun Wen a,b,, Mingzhou Liu a,, Changyi Liu a, Conghu Liu a
a
b

School of Mechanical and Automotive Engineering, Hefei University of Technology, Hefei 230009, China
School of Mechanical and Power Engineering, North University of China, Taiyuan 030051, China

a r t i c l e

i n f o

Keywords:
Remanufacturing
Two-stage uncertain programming
Production planning
Approximation method
Hybrid intelligent algorithm

a b s t r a c t
Remanufacturing is becoming a strategic emerging industry in China. However, there are
many uncertain factors such as remanufacturing rate of recycling products, reprocessing
costs, quantity of recycling products during a remanufacturing process. Hence, it is difcult
to make an accurate production planning. This paper aims at studying a new remanufacturing production planning model in view of some possible uncertain factors in a remanufacturing enterprise according to the features and characteristics of remanufacturing.
Considering the production capacity constraint of recycling, reprocessing and reassembly
under the condition of uncertain reprocessing amount, unpredictable reprocessing cost,
unknown purchase volume of new parts, and uncertain customer demand, this paper
develops a two-stage, multi-period hybrid programming model with compensation function based on uncertainty theory to minimize the total remanufacturing cost. A hybrid
intelligent algorithm is designed combined with compensation function approximation,
neural network training, and virus particle algorithm to optimize this two-stage uncertain
remanufacturing production planning. By use of compensation function approximation
method, it is to convert an innite optimization problem in this algorithm into that of a
nite one. Finally, one remanufacturing simulation case is studied to validate the efciency
and rationality of the proposed approach.
Crown Copyright 2015 Published by Elsevier Inc. All rights reserved.

1. Introduction
Remanufacturing is a recycling process to manufacture a recycling product as good as a new one [1]. Chinas 12th FiveYear Plan for Circular Economy put forward a strategy deployment to build a circular industrial system clearly. As a strategic
emerging industry, remanufacturing is also becoming an effective way to develop circular economy and promote social sustainable development [2]. Just taking automobile as example, it is estimated that the total number of Chinese civilian vehicles will be 200 million in 2015 and the discarded automobiles more than 10 million for the rst time. With the emergency
of discarded industrial products in a large scale and booming of remanufacturing industries, remanufacturing production
planning and scheduling as one of key links in the production management, the study on it is of important theoretical
and practical value for improving production management within a remanufacturing enterprise.

Corresponding authors at: School of Mechanical and Automotive Engineering, Hefei University of Technology, Hefei 230009, China.
E-mail addresses: whjnuc@aliyun.com (H. Wen), liumingzhou0551@163.com (M. Liu).
http://dx.doi.org/10.1016/j.amc.2015.01.070
0096-3003/Crown Copyright 2015 Published by Elsevier Inc. All rights reserved.

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

743

Great stride has been made in the study of remanufacturing production planning optimization. Teunter et al. [3] create a
manufacturing/remanufacturing batch production model when either of the two new and remanufacturing product can
meet the production demand, which aims to minimize the total remanufacturing cost if demand and supply are given. Li
et al. [4] develop a non-constraint, multi-period production planning model considering the possibility of substitution for
remanufacturing products and nd solution using a dynamic programming and heuristic algorithm. Li et al. [5] build a planning model for both manufacturing and remanufacturing hybrid production scheme including emergency procurement. To
achieve the minimum cost, this model is solved by adopting the combination of genetic algorithm and heuristic dynamic
programming method. Doh and Lee [6] establish a remanufacturing production planning model including mixed integer programming to maximize prot based on considering the inuence and constraint of some links such as recycling disposal,
disassembly, inspection, remanufacturing, and reassemble. However, a large number of uncertain factors exist in the remanufacturing production process [7], including: Uncertainty of recycling factors, such as unknown recycling time, variable recycling quantity, and unknown quality; Uncertainty of disassembly rate of recycling product and its component, it means that
it is unknown whether a product or a component can be disassembled or not; Uncertainty of remanufacturing rate of recycling products due to out-of-tolerance, residual stress, internal cracks, surface deformation, and so forth; Uncertainty of customer demand; and so forth. Due to the existence of these uncertain factors, traditional production planning and control
methods cannot adapt to remanufacturing production system, and it is more complicated to make a remanufacturing production planning than a traditional production planning. Therefore, Guide [8] study the inuence on remanufacturing production planning and control because of a random uncertainty in a remanufacturing production system, the imbalance
between recycling supply and demand, and other unknown factors of recycling products. For one-off product and the impact
of a random machine failure, Kenne et al. [9] study a strategy scheme for a mixed production of both manufacturing and
remanufacturing under a closed-loop reverse logistics network in order to reduce the inventory costing. Jing et al. [10,11]
establish a production planning model for a hybrid manufacturing and remanufacturing system under an uncertainty environment, and develop a diploid genetic algorithm to nd solution against the uncertainty of customer demand, recycling
quantity, remanufacturing cost and remanufacturing rate. Wang and Wang [12] study a closed-loop optimal supply chain
decision-making approach from the view of the supply chain when it happens variation in market demand, remanufacturing
cost, as well as recycling cost. Mukhopadhyay [13] studies a remanufacturing production planning system when customer
demand and recycling supply follow random distribution. Su et al. [14,15] create a hybrid uncertain programming model
considering the stochastic and fuzzy parameters in the remanufacturing process taking prot maximization in multi-period
and cost minimization in single period as goal respectively. Huang and Chen [16] construct a hybrid integer programming
model for a fuzzy remanufacturing system and present a crisp equivalent form for this model to minimize the total cost
based on the capacity constraints of recycling, disassembly, reprocessing, and reassembly.
During these studies mentioned above, some remanufacturing production planning models are only suitable for those
remanufacturing production system whose factors related are known, and some could not adjust and compensate the
production plan when the production plan must be changed because of disturbance events although some uncertain factors and disturbance events are taken into account. In this paper, a two-stage remanufacturing production planning
model is established based on uncertainty theory in order to minimize the total cost. This model contains the inuence
of uncertain factors on remanufacturing production planning and regards some uncertain parameters including customer
demand, remanufacturing quantity, purchase volume of new parts, and remanufacturing costs as uncertain variable with
known distribution. It is divided into two stages: The rst stage includes recycling and disassembly, and the second stage
contains inspection, reprocessing, purchasing of new parts, and reassembly. Some decisions have to be made before some
uncertain parameters are determined in the rst stage, so it is allowable that the decision in the rst stage is infeasible,
then, some correspond measurements must be taken to nd optimal scheme for the production plan in the second stage.
Following, the compensation function approximation is studied and a hybrid intelligent algorithm combining compensation function approximation, neural network training, and virus particle algorithm is developed to solve this model. The
purpose of this paper is to provide a new way to make a remanufacturing production planning under uncertain
conditions.
The paper is organized as follows: Section 2 gives a mathematical description of a two-stage uncertain programming.
Then a two-stage uncertain remanufacturing production planning model is established in Section 3. Following, a solution
method of this proposed model is developed in Section 4. Section 5 gives a case study to validate the model and algorithm
proposed in Sections 3 and 4. Finally, Section 6 draws conclusions and suggests the next study direction of this theme.
2. Two-stage uncertain programming denition and its mathematic description
Some decision parameters in the remanufacturing production plan are uncertain, and they are connected with each other
[17,18].
At rst, several denitions are identied as follows:
Denition 1. Let C be a nonempty set, L be one of algebra r of C, and each element K in algebra
function M meets the following three principles when L 2 0; 1:

r is called an event. If

744

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

(1) MfCg 1.
(2) For any event K, MfKg MfKc g 1.
P1
(3) For any sequences with countable events fKi g, Mf[1
i1 MfKi g
i1 Ki g 6
then M is called an uncertain measure, triple C; L; M is called an uncertainty space.
Denition 2. An uncertain variable is a function n from an uncertainty space C; L; M to the set of real numbers such that
fn 2 Bg is an event for any Borel set B, it is:

fn 2 Bg fc 2 Cjnc 2 Bg

Denition 3. Suppose n be uncertain variable, and at least one of integral from the right side of formula (2) is limited (in
order to avoid the appearance of 1  1), then

En

Mfn P rgdr 

Mfn 6 rgdr

1

is called an expected value of uncertain variable n.

Then a programming model with uncertain parameters is constructed as follows:

8
min cT x qT ny;
>
>
>
< s:t: Ax b;
>
Tnx Wny hn;
>
>
:
x P 0; y P 0;

where qn, Tn, Wn and hn are uncertain variables dened in uncertainty space C; L; M. Due to the inuence of uncertain parameters, the model is difcult to solve by conventional methods. So a new approach is: to divide the whole process
into two stages. At rst, to determine the decision variables x and observe the realization of the uncertain variable n, and
declare the decision-making y at the end of the rst phase. Following, to compensate the insufciency caused by the rst
phase because of uncertain factors during the second phase. The compensation function in the second phase is then formed:

8
T
>
< min q ny;
s:t: Wny hn  Tnx;
>
:
y P 0:

Let Q x; n minfqT nyjWny hn  Tnx; y P 0g, Q x; n is called the optimal function of the second stage. To avoid
infeasible solutions, Q x; n is restricted to 1 and 1 1 1. Let K fxjx 2 Rn1 ; CrfQ x; n < 1g 1g, x 2 K is
called induced constraint, which can ensure to nd a feasible solution for each realization in n to solve the compensation
problem. Compensation function Q E x is denoted as Q E x En Q x; n, where En is operator of an expected value for uncertain variable n.
Combined with the compensation function, the two-stage uncertain programming problem can be shown by the
following:

8
min cT x Q E x;
>
>
>
< s:t: Ax b;
>
x P 0;
>
>
:
x 2 K;

8
T
>
< Q x; n min q ny;
s:t: Wny hn  Tnx;
>
:
y P 0:

The aforementioned model contains uncertain variables with unlimited support and could not be solved easily because of
innite dimensional optimization problem, so it is necessary to converted it into nite dimensional optimization problem by
Q
use of the approximation method. Suppose uncertain variable n n1 ; n2 ; . . . ; nr T has unlimited support N ri1 ai ; bi ,
where ai ; bi  is the support of ni , then the possibility distribution of uncertain variable n can be approximated by that of
uncertain variable sequence ffm g, where fm fm;1 ; fm;2 ; . . . ; fm;r T . 8i 2 f1; 2; . . . ; rg, fm;i g m;i ni can be dened as follows:

g m;i ni

8
< ai ;

ni 2 ai ; ai m1 ;
n 
o
ki 
k
: sup m
ki 2 Z; s:t: mi 6 ni ;

ni 2 ai m1 ; bi ;

For 8i 2 f1; 2; . . . ; rg, when ni 2 ai ; bi , the value of uncertain variable fm;i can be obtained by the follows:

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

nr 
i
ri mai ;
m

o
r i ki mai  1; . . . ; K i ;

745

where r is the greatest integer no more than r. If mbi is an integer, then K i mbi  1, otherwise K i mbi . The possibility
n
o
 
v m;i of uncertain variable fm;i is denoted by v m;i mri Pos cj mri 6 ni c < rim1 .
Then, formulas (5) and (6) can be converted into the following style through approximation method:

b m x;
min ^zm x cT x Q

s:t: x 2 D;
b m x Ef Q x; fm , for the specied x and c,
where Q
m

8
T
>
< Q x; fm c min q fm cy;
s:t: Wfm cy hfm c  Tfm cx;
>
:
y P 0:

10

3. Design of a two-stage uncertain remanufacturing production planning model

3.1. Project statement
This paper studies only remanufacturing production planning-making for one kind of product/component with recycling
products. This product may consist of N parts, its production plan involves T identical cycles, and uncertain variables in each
cycle are mutual independent. Remanufacturing work ow is shown in Fig. 1, and its operation principle is as follows:
(1) Recycle discarded products by a third party recycling company, then a remanufacturing enterprise purchases the recycling parts according to the contract from this third party company and forms a recycling product inventory.
(2) Disassemble the recycling products then divide all parts into groups according to material attribution. Finally, after
cleaning and inspection, each part is identied as use directly, remanufacturable, or scrap according to the technical
requirements.
(3) Renew all remanufacturable parts through remanufacturing technology.
(4) Assemble the nished products with remanufacturing parts. If necessary, buy some new parts from suppliers to make
up for the shortage. Finally, store them in the warehouse.
(5) Deliver the nished products to customer.
There are a lot of uncertain factors during the remanufacturing production, even some will be found during the production process gradually. For example, the accurate reprocessing cost and recycling quantity of some constituent parts could be
known only after disassembly and inspection because of their different utilization time, operating environment and damage
degree. However, some decisions must be made before uncertain parameters are identied, so the remanufacturing process
is divided into two phases to nd optimal production planning scheme: An expected value model including uncertain
parameters is built in the rst phase, and it is allowable to make some infeasible decision; In the following second phase,
it will revise the mistakes of the rst phase by use of compensation method after the uncertain parameters are identied.
Meanwhile, this two-stage production planning model still needs to meet the following assumption:

Product recycling
use directly

third party recycling inventory of disassemble wash and test

remanufacturing
manufacturers recycled product

assemble

recycling
scrap

new parts

Fig. 1. Work ow of a remanufacturing production system.

inventory of demand for

remanufactur remanufactu
ed products red products

746

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

(1) The total remanufacturing cost contains recycling cost, disassembly costs, cleaning costs, inspection costs, remanufacturing costs, purchasing costs of new parts, assembly costs, and inventory costs for recycling products and remanufacturing products.
(2) Disassembly, reprocessing and reassembly are done within the constraints of production capacity.
(3) New parts can be free supplied.
(4) Customer demand, amount of new parts, amount of remanufacturable parts and reprocessing cost are uncertain variables in each cycles.
(5) The recycling products must be inspected roughly in the rst phase and it is impossible to discard a product in whole
in the second phase.
3.2. Model study
In order to establish a two-stage uncertain remanufacturing production planning expected value model, it is rstly to
clarify that the indices, parameters and decision variables in it are dened as follows:
(1) Subscript
i, i 1; 2; . . . ; N: indicate the type of a part or a component.
t, t 1; 2; . . . ; T: denote the period number of production plan.
(2) Notations
C R : recycling Cost per unit product.
x1R;t : amount of recycling products in the rst phase at the period t.
x2R;t : amount of recycling products in the second phase at the period t.
C I : inventory cost per recycling product.
x1I;t : inventory of recycling products in the rst phase at the period t.
x2I;t : inventory of recycling products in the second phase at the period t.
C D : disassembly cost per recycling product.
x1D;t : disassembly amount of recycling products in the rst phase at the period t.
x2D;t : disassembly amount of recycling products in the second phase at the period t.
cnit : purchasing cost of part i at the period t.
C A : assembly cost per remanufacturing product.
yA;t : assembly amount of remanufactured products at the period t
C PI : inventory cost of remanufactured products.
yP;t : inventory volume of remanufactured products at the period t.
ED;t : maximum production capacity of disassembly process.
EM;it : maximum production capacity of remanufacturing process for part i.
EA;t : maximum production capacity of reassembly process.
HD;t : disassembly time per recycling product.
HM;it : remanufacturing time for part i.
HA;t : reassembling time per nal product.
ni : amount of part i in a product.
max IR;t : the maximum inventory capacity for recycling products.
max PR;t : the maximum quantity for recycling product.
max IP;t : the maximum inventory capacity for remanufacturing nished products.
(3) Uncertain variables
nm
it c: reprocessing amount of part i at the period t.
nnit c: purchase volume of part i at the period t.
ndt c: customer demand at the period t.
ncit c: reprocessing cost of part i at the period t.
Following, it is to describe how to establish this model.
It is difcult to predict damage degree of each part before a recycling product is dismantled, thus some uncertain variables
such as purchase volume and remanufacturing costs are all unknown, and they can be identied only after disassembly and
inspection. However, some decisions must be made before that, so the remanufacturing production process is divided into
two phases to optimize such kind of production planning. x1R;t ; x1I;t ; x1D;t is called the decision vectors in the rst phase, i.e.
n
c
d
1
1
1
x1R;t ; x1I;t ; x1D;t must be determined before the realization of uncertain variable nm
it c; nit c; nt c; nt c and xR;t ; xI;t ; xD;t is also
allowed to be infeasible. In the second phase, an enterprise can continue purchasing the recycling products from the third party
recycling company if remanufacturing nished products cannot meet customer demand and compensate the insufcient estimation in the rst stage. For convenience, we use x1 and n to denote decision vectors x1R;t ; x1I;t ; x1D;t and uncertain variables

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

747

n
c
d
nm
it c; nit c; nt c; nt c respectively. To minimize the total cost, the planning model in the rst phase is constructed as
follows:

8
T 

X
>
>
>
min TC
C R x1R;t C I x1I;t C D x1D;t En Q x; n;
>
>
>
>
t1
>
>
>
< s:t: x1 6 max IR;t ;
I;t
>
x1R;t 6 max PR;t ;
>
>
>
>
>
>
x1D;t Hd;t 6 ED;t ;
>
>
>
:
x1R;t ; x1I;t ; x1D;t P 0;

11

The total cost in the rst phase involves recycling costs, inventory costs of recycling products, disassembly costs and
remanufacturing costs expected value of the second phase.
To give the decision vector x1 and uncertain variables n of the rst stage a realization value nc, the programming model
of the second phase can be expressed as the following form:

Q x; nc min

T 
T
T X
N
 X
X
X
 m

C R x2R;t C I x2I;t C D x2D;t
C A yA;t C PI yP;t
nit cncit c nnit ccnit ;
t1

t1

12

t1 i1

The total cost in the second phase mainly includes three sections: the compensation value for the rst stage
C R x2R;t C I x2I;t C D x2D;t , sum of assembly costs and inventory costs for nished products, and sum of reprocessing costs and
new parts purchasing costs, where reprocessing cost, amount of reprocessing part, and purchase volume of new part are
all uncertain variables.



s:t: x1D;t x2D;t HD;t 6 ED;t

t 1; 2; . . . ; T;

nm
it cH M;it 6 EM;it ; i 1; 2; . . . ; N;

13

t 1; 2; . . . ; T;

14

yA;t HA;t 6 EA;t ; t 1; 2; . . . T;

15


 

x1I;t  x1I;t1 x2I;t  x2I;t1 x1D;t x2D;t x1R;t x2R;t 6 max PR;t ;
yA;t yP;t1  yP;t P ndt c;

t 1; 2; . . . ; T;



n
x1D;t x2D;t ni P yA;t ni P nm
it c nit c;
x1D;t x2D;t P yA;t ;

yP;t 6 max IP;t ;

x1R;t ;

x1I;t ;

x1D;t ;

i 1; 2; . . . ; N;

x2D;t ;

18

20

t 1; 2; . . . ; T;
x2I;t ;

t 1; 2; . . . ; T;

19

t 1; 2; . . . ; T;

x2R;t ;

16
17

t 1; 2; . . . ; T;

x1I;t x2I;t 6 max IR;t ;

t 1; 2; . . . ; T;

21
yA;t ;

yP;t P 0;

22

Formulas (13)(15) denote the production capacity constraints of disassembly process, remanufacture machining process
and remanufacture assembly process respectively; formula (16) shows that amount of recycling products is less than the
maximum recycling capacity at the period t; formula (17) denotes that remanufacturing product quantity is more than customer demand; formula (18) shows that amount of part i after disassembly should not be less than the reassembly amount
and the reassembly quantity should not be less than the sum of new-buying parts and the qualied remanufacturing parts;
formula (19) means that disassembly amount should not be less than that of reassembly; formulas (20) and (21) show the
inventory capacity constraints of recycling products and remanufacturing products respectively; formula (22) expresses the
non-negative constraints of variables.
Suppose S be a set of collection of x that meets restraints of formula (11), for some x 2 S, the model solution of the second
phase may be nonexistent. Here, it is to add induced constraints K fxjCrfcjQ x; nc < 1g 1g on decision vector of the
rst phase x 2 S, the feasible region of model is K \ S. Corresponding to the remanufacturing production planning model, the
decision vector is feasible if the followed condition is met:

ni

T
X
t1

max PR;t P ni

T
X
t1

max ndt c 

T
X

min nnit c:

23

t1

The maximal quantity of recycling products of a part obtained by remanufacturing production should not be less than the
difference between the maximal total demand and the minimal purchasing quantity.

748

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

4. Solution method
4.1. Approximation algorithm of expectation function
Since n is a continuous uncertain variable with innite support in this two-stage uncertain production planning model, it
is impossible to evaluate the precise value of En Q x; n at point x. In this paper, it is to convert innite dimensional optimization problem into a nite one by taking an approximation strategy.
n
c
d
m
n
c
For a continuous uncertain variable n nm
it c; nit c; nt c; nt c, let fm g m n g m;1 nit c; g m;2 nit c; g m;3 nt c;

g m;4 ndt c be a discrete uncertain variable. f x En Q x; n is approximated to f m x Efm Q x; fm , for 8m,











b
f qm b
f qm;1 ; b
f qm;2 ; b
f qm;3 ; b
f qm;4 q 1; 2; . . . ; L, where L L1 L2 L3 L4 . Let v q v m;1 b
f qm;1 ^ v m;2 b
f qm;2 ^ v m;3 b
f qm;3 ^ v m;4 b
f qm;4 , v m;i
is the possibility distribution of uncertain variable fm;i . The optimum value of an uncertain remanufacturing production
planning of the second stage is expressed by the following formula:



g q x Q x; bf qm c
min

T 
T
T X
N 
 X

X
X
bf q cbf q c bf q ccn :
C R x2R;t C I x2I;t C D x2D;t
C A yA;t C PI yP;t
it
m;1
m;3
m;2
t1

Without loss of generality, the subscript of

denoted as below:

wq
where

t1

vq

24

t1 i1

and g q x are rearranged, and meeting g 1 x 6 g 2 x 6    g L x, wq is





q
q1
L
L1
1
1
maxv p  maxv p
maxv p  max v p ;
pq1
p0
2 p1
2 pq

25

v 0 v L1 0. When the decision vector is x, f m x can be calculated as follows:

f m x

L
X
wq g q x;

26

q1

When m is large enough, expectation function f x of a remanufacture production planning in the second stage can be
approximated by use of formula (26). Then the uncertain production planning model can be represented through the following approximation algorithm:

min TC

T 

X
C R x1R;t C I x1I;t C D x1D;t f m x:

27

t1

4.2. Neural network training

In view of the complexity of the model mentioned above, this paper selects a three layer radial basis function neural network (RBFNN) to replace formula (26) in order to accelerate the solution tempo. RBFNN belongs to the feedforward neural
network and has a simple structure, strong approximation ability and rapid learning convergency, so it is applicable to problems such as pattern recognition, function approximation and adaptive ltering. When used in function approximation,
RBFNN adopts nonlinear activation function as basis function of hidden layer nodes and linear function for output nodes.
RBF is radial symmetry, so for neurons, the farther the input value from the center, the lower the degree of activation, i.e.
local feature.
4.3. A virus particle swarm algorithm
Uncertain optimization with compensation is a nonconvex optimization problem and the model is complex and difcult
to solve. The application of particle swarm optimization (PSO) helps to nd the satised solution in the feasible time. But a
large number of engineering practice shows that the optimization efciency is low and the global optimization ability is poor
for the basic particle swarm algorithm. This paper introduces the virus evolution principles based on PSO algorithm to
strengthen the information exchange between the same generation particles, enhance the population diversity and improve
the precision and convergence speed against these weaknesses. Virus particle swarm optimization algorithm [19] (VPSO) is
adopted to solve the uncertain production planning problem. Two groups are generated in the process of evolution: main
groups and viruses groups. The main groups correspond to the solution space of problem and each particle represents a feasible solution. Main groups are responsible for communication longitudinally between generations to make the population
global optimization.
Part of the chromosome gene of virus individual is changed to change their genetic information by transcribing gene segment of some individual and transcribing reversely to other individual. Virus transfers evolutionary information crosswise
among individuals of same generation and conducts the local search of particle swarm.

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

749

4.3.1. Algorithm parameter settings

Fitness function of main groups: according to the principle of the minimum total production cost, formula (27) is taken as
the tness function of main groups.
Infection depth: The infection depth of virus individual to main groups is measured by the tness function of virus
individual.

fitv irusj

S
X
FithostPrej;k  FithostAftj;k ;

28

k1

where fit v irusj is the tness of virus j, S is the set of main individual infected by virus j, FithostPrej;k and FithostAft j;k represent
the tness of kth individual before and after infected by virus j respectively.
Virus vitality: The vitality of virus j in t + 1 generation is dened as:

LF j;t1 z  LF j;t fitv irusj ;

29

where z is the rate of life attenuation.

Virus activity: The probability that a virus is selected as infection subject during evolution

PC j

1
1

m ;
1 aefitv irusj 1 be

30

where m is the virus number that has the same gene encoding with virus j. a and b are weighting coefcients used to balance
the activity of virus individual and group. Generally a > b is met to increase inuence of the activity of virus individual on the
selection probability. Bigger probability indicates the stronger virus individual vitality and the greater individual tness
value. Yet the excessive probability will reduce population diversity and converge until local optimal. Too small probability
will reduce the convergence speed.
Virus update mechanism: whether the virus individual updates is measured by the similarity of individual chromosome.
The similarity between individual k and j is represented as:

Hjk

1
; djk 2 0; 1;
1 djk

where djk is called the chromosome euclidean distance between individual k and j, djk

31
q

Pn
2
l1 Bjl  Bkl , n is the chromo-

some genes quantity, Bjl and Bkl are the lth chromosome gene of individual j and k respectively. When Hjk reaches the given
standard H, new virus individual will be generated in the next generation to ensure the diversity of population.
Encoding scheme: The papulation is made up by pop  size particles, for a two-stage remanufacturing production planning model, the search space of decision variables is 3T dimension, so using X xR;1 ; . . . ; xR;T ; xI;1 ; . . . ; xI;T ; xD;1 ; . . . ; xD;T
denotes the decision vectors in the rst stage. Virus particles in the group are taken as the code substring of the main individual, so it is to select a variable of a random period from the decision about recycling quantity, inventory amount and disassembly quantity of recycling products, xR;u ; xI;v ; xD;w , where u; v ; w 2 1; T.
4.3.2. Calculation procedure
Step 1. Parameter initialization. Recycling quantity, inventory and disassembly quantity of each period should not exceed
the maximal number of recyclable products because most of remanufacturing production are addressed with a customer order. Therefore, it can generate a main group particle and corresponding virus individual random from
0; max PR;t , then validate the feasibility of decision vector X. If it meets formula (23), it is feasible and can be taken
as a particle of initial population. Repeating above-mentioned process until to obtain pop  size feasible particles of
main group X 1 ; X 2 ; . . . ; X popsize .
Step 2. Use formula (26) again to obtain the approximate compensation value of every particle X j , then set the particles
position and the target as their own optimal position and optimal value, denoted by pbest. Find out the optimal particle in the population and set its optimal position and target, denoted by gbest. Set V be the upper bound of particle
velocity and value V the difference of maximum quantity and minimum quantity of recycled products.
Step 3. Each particle updates its speed and position in the solution space according to the formula of a basic particle swarm
optimization.

v ti wv t1
c1 rand1 pi  X t1
c2 rand2 pg  X t1
;
i
i
i

32

X ti X t1
v ti ;
i

33

where i 1; 2; . . . ; pop  size, w is called inertia coefcient, c1 and c2 are learning factors, rand1 and rand2 are random
numbers distributed evenly in the range of [0, 1]. The value of w can be determined by the followed formula:

w wmax 

wmax  wzmin
 iter;
itermax

34

750

Step
Step
Step
Step

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

where wmax , wmin are the maximum and minimum of w respectively; iter and itermax are the current iteration number
and maximum number of iterations respectively. The feasibility of every particle X ti can be veried by use of constraint
formula (23). If feasible, if is treated as a new particle; otherwise, it is to regenerate through (32) and (33).
4. Virus infection. Virus individual j infects the main group particles xk randomly to generate another main individual xl
through selection probability PC j .
5. Recalculate the tness value of each main group particle and update pbest and gbest. The tness value, vitality and
activity of virus individual will be calculated according to the tness before and after being infected.
6. t t 1.
7. Termination condition of algorithm: The algorithm will be terminated when the preset maximum number of iterations is reached or some acceptable solutions come true. Otherwise, turn to step 3 to repeat this calculation.

5. Case study
In order to validate this model and algorithm, this paper takes a following case as example. A remanufacturing production
planning for 4102QB diesel engine for 3 months is made using the proposed approach. Considering the complex structure
and numerous parts of engine, this paper only takes engine camshaft assembly. The camshaft assembly parts is shown in
Fig. 2 and their parameters in Table 1.
Suppose that the recycling product inventory and the nished products inventory are zero at the beginning of the period.
The other parameters are as follows: C R 60, C I 2, C D 5, C A 17, C PI 4; max IR;t 210, max IP;t 280;
ED;t 29000 min, EA;t 35000 min; HD;t 90 min, HA;t 110 min; For the sake of simplicity, the change of the time dimension of these parameters is not taken into account in this example. Remanufacturing processing capability and operation
time are as shown in Table 2.
Furthermore, remanufacturing amount, purchase volume, market demand and remanufacturing cost per month are all
normal distribution N e; r. The concrete parameters are as shown in Table 3.
Using the approximation method above mentioned, it generates 3000 samples b
f qm , q 1; . . . ; 3000. Put each sample into
f qm c. Repeat the above-mentioned process and nd a
formula (24) and calculate the optimal value in the second stage Q x; b
set of input and output data for goal function. A RBF neural network is trained to approximate the goal function through the
specied training data (the neuron number of input layer is 5T 2TN = 45, the hidden layer is 50, output layer is 1), then
embed the trained neural network into VPSO algorithm to search the optimal solution. Parameter Settings of VPSO algorithm
are as follows: Inertia coefcient linear decreases from 0.8 to 0.3; Learning factors c1 and c2 is 2; pop size 100; Virus life
attenuation rate z 0:8; Weighted coefcient of virus individual activity a 0:6; Weighted coefcient of virus group activity
b 0:4; Number of iterations iter = 3000. The algorithm is programmed in MatlabR2009 under the computer system
required as follows: Intel i3CPU, 2.3 GHZ processor, 4 GB RAM. The algorithm is run independently replicated 10 times.
The optimal solution of remanufacturing production planning model in the rst stage is (310, 0, 310, 296, 15, 281, 234,
27, 249), the optimal solution in the second stage is (0, 0, 0, 290, 10, 36, 0, 36, 276, 21, 0, 0, 0, 293, 10), the corresponding
optimal value of the total remanufacturing cost is 175664.2.
To illustrate the optimization performance of VPSO algorithm, this paper adopts PSO and VPSO to solve the two-stage
uncertain remanufacturing production planning model under the same initial conditions. The results is shown in Table 4,
where BF, AF, and MI represent the iterative optimal solution, average optimal solution after 10 cycles, and the largest number of iterations respectively.
The results show that the lower total cost value can be got through VPSO than by PSO algorithm. Because VPSO imports
the virus evolution infection function, the population diversity during evolution process especially at its late period is
ensured by virus individual infecting main groups. And VPSO also avoids the prematurity of the population and improves
the solution precision of the algorithm. Iterative times of VPSO algorithm is less than that of PSO algorithm, although the

12

1
2

13

6
8 7
9
10
0

11

Fig. 2. Graphical representation of camshaft assembly.

751

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

Table 1
Camshaft assembly parts.
No.

Parts code

Parts name

Remanufacturable*

Number

New product price ()

4100QB-01-014-II
4100QB-01-014
4102QB-02-001
GB1096-79
GB93-87
GB21-76
4100QB-01.02-001
4100QB-01.02-002
GB93-87
GB21-76
4100QB-02-004
4100QB-01-013A
4100QB-2-02-006
495QA-02-005

Camshaft bush II
Camshaft bush III
Camshaft
Flat key A8  22
Washer 8
Bolt M8  25
Camshaft idle gear
Idle Gear
Washer 12
Finished hexagon head bolt M12  32
Camshaft thrust plate
Camshaft bush I
Valve pushrod
Valve tappet

N
N
Y
N
N
N
Y
N
N
N
Y
N
Y
Y

1
1
1
1
2
2
1
1
1
1
1
1
8
8

10.3
6.4
197.5
2.5
0.8
1.5
62.5
1.4
0.8
1.4
14.5
7.6
12.2
13.4

2
3
4
5
6
7
8
9
10
11
12
13
*

For security guarantee, all fasteners and wearing parts should be forced to scrap whatever damaged or not.

Table 2
Remanufacturing processing capacity constraints (min).
Parts

Camshaft

Camshaft idle gear

Camshaft thrust plate

Valve pushrod

Valve tappet

EM;it
HM;it

16,000
45

15,000
35

5000
15

50,000
5

75,000
10

Table 3
Parameters of remanufacturing plan model.
Parts name

Uncertain variable

Camshaft

nm
it c
nnit c
ncit c

(230, 20)
(20, 3)
(50, 10)

(275, 15)
(70, 20)
(25, 10)

(249, 10)
(38, 6)
(25, 5)

Camshaft idle gear

nm
it c
nnit c
ncit c

(256, 11)
(43, 13)
(34, 14)

(250, 15)
(50, 15)
(25, 5)

(170, 15)
(134, 11)
(20, 8)

Camshaft thrust plate

nm
it c
nnit c
ncit c

(132, 12)
(25, 3)
(5, 3)

(148, 5)
(26, 5)
(6, 3)

(165, 9)
(25, 4)
(7, 3)

Valve pushrod

nm
it c
nnit c
ncit c

(632, 62)
(342, 30)
(1.5, 0.4)

(445, 31)
(530, 80)
(2.1, 0.5)

(525, 58)
(480, 20)
(2.5, 0.6)

Valve tappet

nm
it c
nnit c
ncit c

(914, 50)
(255, 21)
(3.1, 0.4)

(580, 30)
(368, 24)
(1.9, 0.3)

(450, 45)
(240, 25)
(2.4, 0.3)

Camshaft assembly

ndt c

(280, 20)

(300, 10)

(230, 30)

Table 4
Comparison of optimized results by PSO and VPSO.
Optimization algorithm

BF

AF

MI

PSO
VPSO

180368.6
175664.2

182547.4
176734.5

3180
2650

former is more complex than the latter. The reason is that VPSO algorithm is not the simple superposition of PSO and virus
mechanism, but the optimized combination of both. The speed of the genetic mutations caused by virus infection is dozens
of time orders of magnitude faster than that on their own and its inuence is dozens of orders of magnitude larger. The algorithm can avoid population trapping in local optimum and improve the convergence speed of VPSO at the same time.
For any given feasible decisions, approximate expectation function f m x is a sequence fam g about number m of sample.
When m tends to innity, the sequence fam g converges to a0 . The limitation a0 is the value of expectation function corre-

752

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

approximate
value of compensation function

110900
110800
110700
110600
110500
110400
110300

500

1000

1500

2000

2500

3000

sample size
Fig. 3. The convergence of approximate expectation function f m x.

sponding to decision x, so f m x converges to expectation function f x when m tends to innity. This situation can be seen
from Fig. 3. To select an optimal solution of decision variable x = (310, 0, 310, 296, 15, 281, 234, 27, 249), then the value of
f m x is 110876.11. Observing Fig. 3, the limiting value of f m x located in [110,800, 110,900]. So the relative deviation of f m x
and a0 are small. The results show that this hybrid algorithm to solve this two-stage uncertain remanufacturing production
planning problem is effective.
6. Conclusion
A production planning is one of important job of a remanufacturing production system. However, it is difcult for dispatcher and scheduler to make efcient production planning because of the inuence of large amount of uncertain factors,
and such production planning leads to the chaos of production and low economic benet generally. Hence, it studies a
remanufacturing production planning scheme under uncertain conditions in this paper. Its main contents are as follows:
(1) To establish a new two-stage remanufacturing production planning expected value model based on uncertainty theory and two-stage uncertain optimization method in view of the uncertainty of remanufacturing production. Some
parameters must be given expected value although they are unknown in the rst stage, and some measurements must
be taken to compensate the difference between the estimation in the rst stage and the actual demand in the second
stage, which aims at getting close to the actual production requirement.
(2) To study the approximation method of expectation function and to develop a hybrid intelligent algorithm combined
approximation method, neural network training, and virus particle algorithm in order to nd solution for this twostage remanufacturing production planning model. This algorithm can convert an innite optimization problem into
that of a nite one efciently, and it can ensure the accuracy of solution through virus particle algorithm and accelerate the solution tempo.
(3) To study a case in order to verify the feasibility and effectiveness of the proposed planning model and algorithm. The
result shows that this model and algorithm can optimize remanufacturing production planning-making under some
unknown factors.
Despite some uncertain factors in the remanufacturing production is considered, it is insufcient to mirror the whole
actual remanufacturing production environment. Hence, the next study will add the impact on remanufacturing production
planning caused by recycling amount, recycling product quality, and so on. In addition, the study on a remanufacturing production planning for diversied products will be the direction in the future.
Acknowledgements
This work was supported by Major Project of National Basic Research Program of China (No. 2011CB013406).
References
[1] B. Xu, Remanufacture engineering and its development in China, Chin. Surf. Eng. 23 (2010) 16.
[2] B. Xu, S. Dong, P. Shi, States and prospects of China characterised quality guarantee technology system for remanufactured parts, J. Mech. Eng. 49
(2013) 8490.
[3] R.H. Teunter, Z.P. Bayindir, W.V. Den, Heuvel, Dynamic lot sizing with product returns and remanufacturing, Int. J. Prod. Res. 44 (2006) 43774400.
[4] Y. Li, J. Chen, X. Cai, Uncapacitated production planning with multiple product types, returned product remanufacturing and demand substitution, OR
Spectr. 28 (2006) 101125.

H. Wen et al. / Applied Mathematics and Computation 256 (2015) 742753

753

[5] Y. Li, J. Chen, X. Cai, Heuristic genetic algorithm for capacitated production planning problems with batch processing and remanufacturing, Int. J. Prod.
Econ. 105 (2007) 301317.
[6] H.H. Doh, D.H. Lee, Generic production planning model for remanufacturing systems, Proc. Inst. Mech. Eng., Part B: J. Eng. Manuf. 224 (2010) 159168.
[7] J. Xie, Z. Zhao, L. Kong, et al, An affecting factors and optimizing model for the remanufacturing production planning, Syst. Eng. 25 (2003) 5359.
[8] J. Guide, Production planning and control for remanufacturing: industry practice and research needs, J. Oper. Manage. 18 (2000) 467483.
[9] J.P. Kenne, P. Dejax, A. Gharbi, Production planning of a hybrid manufacturingremanufacturing system under uncertainty within a closed-loop supply
chain, Int. J. Prod. Econ. 135 (2012) 8193.
[10] Y. Jing, X. Wang, W. Li, et al, Optimization for lot sizing production planning with product remanufacturing and demand substitution, Chin. Mech. Eng.
23 (2012) 21762181.
[11] Y. Jing, X. Wang, W. Li, et al, Hybrid lot sizing production planning of manufacturing and remanufacturing to automobile engine, Comput. Integr.
Manuf. Syst. 19 (2013) 774781.
[12] Y. Wang, X. Wang, Research on contract coordination and price decision of closed-loop supply chain under disruption condition, J. Mech. Eng. 49
(2013) 185192.
[13] S.K. Mukhopadhyay, H. Ma, Joint procurement and production decisions in remanufacturing under quality and demand uncertainty, Int. J. Prod. Econ.
120 (2009) 517.
[14] C. Su, Z. An, Remanufacturing production planning considering quality cost and demand substitution, J. Southeast Univ. 27 (2011) 222226.
[15] C. Su, Y. Sha, Remanufacturing production planning based on mixed-uncertainty and evidence theory, J. Southeast Univ. Nat. Sci. Ed. 40 (2010) 712
716.
[16] W. Huang, W. Chen, Remanufacturing production planning based on credibility theory in fuzzy environments, Ind. Eng. Manage. 17 (2012) 2126.
[17] Y. Liu et al, Fuzzy Optimization Method and Application, Science Press, Beijing, 2013.
[18] B. Liu, Uncertainty Theory, second ed., Springer Verlag, Berlin, 2007.
[19] C. Ji, W. Xie, X. Zhu, et al, Optimized load allocation of cascaded hydropower stations based on virus particle swarm optimization algorithm, J.
Hydroelectric Eng. 31 (2012) 3843.