Journal of Applied Mathematics and Computing (2022) 68:2173–2186

https://doi.org/10.1007/s12190-021-01607-9

ORIGINAL RESEARCH

The due date assignment scheduling problem with the

deteriorating jobs and delivery time

Jin Qian1 · Haiyan Han1

Received: 9 May 2021 / Revised: 13 July 2021 / Accepted: 14 July 2021 /

Published online: 3 August 2021
© Korean Society for Informatics and Computational Applied Mathematics 2021

Abstract
This paper considers the single machine scheduling problem with three different due
dates in which the actual processing time of the job is a simple deterioration function
of the starting time. The goal is to minimize the total costs that contain the earliness,
tardiness and due date. We prove that these problems are polynomial time solvable,
and we propose the corresponding algorithms to obtain the optimal sequence and due
date.

Keywords Scheduling · Deteriorating job · delivery time · Common due date · Slack
due date · Different due date · Earliness · Tardiness

1 Introduction

In actual production scheduling, it is an important phenomenon that jobs have due

dates. If the job is completed earlier than the due date, it will have earliness costs. On
the contrary, the job is completed later than the due date, it will have tardiness costs. In
1955, the concept of the due date was first proposed by Jackson [1]. In 2010, Wang et
al. studied the scheduling problem with slack due date and modifying activity [2]. In
2011, Li et al. studied the scheduling problem with deteriorating jobs and due date [3].
In 2014, Yin et al. studied the scheduling problem with common due date, modifying
activity and batch delivery [4]. In 2014, Zhao et al. studied the scheduling problem
with due date and deteriorating jobs [5]. In 2016, Wang et al. studied the scheduling
problem with two agents and due date [6]. In 2017, Li et al. studied the scheduling
problem with common due date and cumulative deterioration effect [7]. In 2017, Liu
et al. studied the scheduling problem with modifying activity, common due date and
linear deteriorating jobs [8]. In 2018, Karhi et al. studied the scheduling problem with

B Jin Qian
[email protected]

1 Department of Mathematics, Northeastern University, Shenyang 110819, People’s Republic of China

123
2174 J. Qian, H. Han

due date and resource [9]. In 2021, Kim et al. studied the scheduling problem with
common due date and improving rate [10].
Generally speaking, the processing time of the job is a constant in most studies. As
the machine wears in many practical cases, the processing time of the job will become
longer. The processing time of the deteriorating job is one of the cases, i.e. the actual
processing time of the job increases with the delay of the starting time. Gupta first
studied the scheduling problem with the deteriorating jobs [11]. In 1994, Mosheiov
first proposed the simple linear deterioration [12]. In 2010, wang et al. studied the
scheduling problem with deteriorating jobs and learning effect [13]. The objective
function is to minimize the sum of due date, earliness and tardiness. They proposed
the polynomially algorithm to solve this problem. In 2011, Yang et al. studied the
flow shop scheduling problem with simple linear deterioration [14]. In 2012, Bai et al.
studied the group scheduling problem with deteriorating jobs and learning effect [15].
In 2013, Wang et al. studied the scheduling problem with deteriorating jobs and convex
resource allocation [16]. In 2015, Yue et al. considered the single machine scheduling
problem with the common due window and deteriorating jobs [17]. In 2017, Azadeh
et al. studied the scheduling problem with deteriorating jobs and learning effect [18].
In 2018, Zhang et al. studied the single-machine scheduling problems with a learning
effect matrix [19]. In 2020, Wu et al. studied a branch-and-bound algorithm and
four metaheuristics for the flow-shop scheduling problems [20]. In 2020, Liang et
al. studied the scheduling problem with the resource allocation and deteriorating jobs
[21]. The objective function was to minimize the weighted sum of completion time and
resource allocation. In 2021, Wu et al. studied the metaheuristics for two-stage flow-
shop assembly problem with a truncation learning function [22]. In 2021, Wu et al.
studied the hyper-heuristic for a single-machine scheduling with two-scenario-based
dependent processing times [23].
In the field of production sequencing and scheduling, the scheduling problem with
the delivery time has received a lot of attention. When a job has been processed, it needs
to be delivered to the customer. This extra time is called past-sequence-dependent
(psd) delivery time. In 2008, Koulamas et al. proposed the concept of past-sequence-
dependent delivery time [24]. In 2013, Liu studied the parallel machine scheduling
problem with learning effect and delivery time [25]. In 2013, Yin et al. studied the
single machine scheduling problem with deteriorating jobs and delivery time [26]. In
2016, Wu et al. studied the single machine problem with learning effect and delivery
time [27]. In 2020, Maecker et al. studied the parallel machine problems with delivery
time [28]. In 2021, Mönch et al. studied the parallel machine scheduling problem with
the delivery time [29].
In this paper, we study the single-machine scheduling problem with delivery time
and deteriorating jobs under three forms of due dates. The actual processing time of
job is a simple deterioration function of the starting time. The objective function is to
minimize the earliness, tardiness and due date. Several properties are proposed. We
propose the polynomial time algorithms to obtain the optimal sequence and due date,
and the complexity of the algorithms are O(nlogn). This paper belongs to the areas
of discrete optimization, theory of computation and theory of algorithms. The rest of
this paper is organized as follows. The problem is described in the second section. The
property is given in the third section. The proofs of the polynomial time algorithm

123
The due date assignment scheduling... 2175

are given from the forth section to the sixth section. The summary is given in the last
section.

2 Notation and problem statement

Suppose there will be n independent jobs S = {J1 , · · · Jn } processed on a machine.

All jobs can be processed at time t0 (t0 > 0). The jobs are processed continuously.
The machine can only process one job at a time. The actual processing time p j of the
job J j is a simple linear deterioration function of the starting time, i.e. p j = b j t j ,
where t j is the starting time of J j , b j is the deterioration rate of J j . The delivery time
q j of the job J j is a simple linear function of the starting time, i.e. q j = r t j , where
t j is the starting time of J j , r is the delivery rate. This paper considers three different
forms of due dates that contain common due date, slack due date and different due
date. They are abbreviated as CON, SLK and DIF respectively. For the common due
date model, all jobs have the same due dates, d j = d, d is the decision variable. For
the slack due date model, the due date d j of job J j is the sum of processing time and
certain parameter q, d j = p j + q, q is the decision variable. For the different due date
model, each job has its own due date d j , d j is the decision variable. The subscript [ j]
indicates that the job is arranged at the jth position in the sequence.
The goal is to minimize the total costs of three parts that are the earliness, tardiness,
and due date (or certain parameter). By the three-region notation [30], the models can
be defined as

n
1| p j = b j t, q psd , C O N | (α E j + βT j + θ d),
j=1


n
1| p j = b j t, q psd , S L K | (α E j + βT j + θq),
j=1


n
1| p j = b j t, q psd , D I F| (α E j + βT j + θ d j ),
j=1

where q psd represents the past-sequence-dependent delivery time; α, β, θ represent

the unit cost of earliness, tardiness, and due date (certain parameter).

3 Preliminary result

Lemma 3.1 In a given job sequence, the positions of any two jobs are swapped and the
positions of other jobs remain unchanged. Then the completion time of the job that is
completed later in two jobs remains unchanged. The starting time and completion time
of the job remain unchanged before the two jobs, and the starting time and completion
time of the job remain unchanged after the two jobs.

Proof Suppose that the sequence of jobs is S1 = {J1 , · · · , Jg , · · · , Jk , · · · , Jn }.

The job Jg is at the ith position, and the job Jk is at the jth position, 1 ≤ i <

123
2176 J. Qian, H. Han

j ≤ n. When the job Jg and the job Jk are swapped, the sequence of jobs is
S2 = {J1 , · · · , Jk , · · · , Jg , · · · , Jn }. The job Jk is at the ith position, and the job
Jg is at the jth position. Let the starting time of Jg be t in the S1 . Then

C S1 (i) = t + bg t + r t = (1 + bg + r )t;
C S1 (i+1) = (1 + bg + r )t + b[i+1] (1 + bg + r )t + r (1 + bg + r )t
= (1 + bg + r )(1 + b[i+1] + r )t;

···

C S1 ( j) = (1 + bg + r )(1 + b[i+1] + r ) · · · (1 + b[ j−1] + r )(1 + bk + r )t.

where C S1 (i) represents the completion time of the job at the ith position in the S1 . In
the S1 , the completion time of Jk is (1 + bg + r )(1 + b[i+1] + r ) · · · (1 + b[ j−1] +
r )(1 + bk + r )t.
The starting time of Jk in the S2 is the same as the starting time of Jg in the S1 .
Then
C S2 (i) = t + bk t + r t = (1 + bk + r )t;

C S2 (i+1) = (1 + bk + r )t + b[i+1] (1 + b j + r )t + r (1 + b j + r )t
= (1 + b j + r )(1 + b[i+1] + r )t;

···
C S2 ( j) = (1 + bk + r )(1 + b[i+1] + r ) · · · (1 + b[ j−1] + r )(1 + bg + r )t.
where C S2 (i) represents the completion time of the job at the ith position in the S2 . In
the S2 , the completion time of Jg is (1+bk +r )(1+b[i+1] +r ) · · · (1+b[ j−1] +r )(1+
bg + r )t. So the completion time of job Jk in the S1 is the same as the completion time
of job Jg in the S2 .
When the job Jg and the job Jk are swapped, it will not affect the starting time and
completion time of the job before the two jobs. Because the completion time of the
job that is completed later in two jobs remains unchanged, it will not affect the starting
time and completion time of the job after the two jobs. 



n
4 The problem 1|pj = bj t, qpsd , CON| (˛Ej + ˇTj + d)
j=1

Lemma 4.1 For any job sequence, the due date d of the optimal scheduling is the
completion time of some job or t0 .

Proof Suppose that the due date d of the optimal scheduling isn’t the completion time
of some job or t0 , i.e. C[h] < d < C[h+1] , 1 ≤ h < n, C[0] = t0 . The objective
function

123
The due date assignment scheduling... 2177


n 
h 
n
Z= (α E [ j] + βT[ j] + θ d) = α(d − C[ j] ) + β (C[ j] − d) + nθ d
j=1 j=1 j=h+1


h 
n
= −α C[ j] + β C[ j] + [hα + nθ − (n − h)β]d.
j=1 j=h+1

(a) When d moves left to C[h] , the objective function is Z 1 , d − C[h] = x > 0.


h−1 
n
Z 1 = −α C[ j] + β C[ j] + [(h − 1)α + nθ − (n − h)β]C[h] ,
j=1 j=h+1

Z − Z 1 = [hα + nθ − (n − h)β]x.
(b) When d moves right to C[h+1] , the objective function is Z 2 , C[h+1] − d = y > 0.


h 
n
Z 2 = −α C[ j] + β C[ j] + [hα + nθ − (n − h − 1)β]C[h+1] ,
j=1 j=h+2

Z − Z 2 = −[hα + nθ − (n − h)β]y.
When hα + nθ − (n − h)β ≤ 0, Z ≥ Z 2 ; otherwise, Z > Z 1 . So d of the optimal
scheduling is the completion time of some job or t0 . 


Lemma 4.2 For any job sequence, the due date d of the optimal scheduling is the
completion time C[h] of the job, h =  n(β−θ)
α+β .

Proof When d = C[h] , the objective function


h−1 
n
Z = −α C[ j] + β C[ j]
j=1 j=h+1
+ [(h − 1)α + nθ − (n − h)β]C[h] .

(a) When d moves left to C[h−1] , the objective function


h−2 
n
Z 1 = −α C[ j] + β C[ j] + [(h − 2)α + nθ − (n − h + 1)β]C[h−1] .
j=1 j=h

Because h is the optimal position,

Z − Z 1 = [(h − 1)α + nθ − (n − h + 1)β](C[h] − C[h−1] ) ≤ 0,

n(β − θ )
h≤ + 1.
α+β

123
2178 J. Qian, H. Han

(b) When d moves right to C[h+1] , the objective function


h 
n
Z 2 = −α C[ j] + β C[ j] + [hα + nθ − (n − h − 1)β]C[h+1] .
j=1 j=h+2

Because h is the optimal position,

Z − Z 2 = [hα + nθ − (n − h)β](C[h] − C[h+1] ) ≤ 0,

n(β − θ )
h≥ .
α+β
n(β−θ) n(β−θ)
So the best position h satisfies α+β ≤h≤ α+β + 1, h =  n(β−θ)
α+β . 


For the convenience of proof, we define two sets: A = {J j ∈ π |C j ≤ d}, B = {J j ∈

π |C j > d}, π is the job sequence.
Lemma 4.3 In the optimal scheduling, the jobs of set A are arranged in descending
order of bi .

Proof There are two adjacent jobs Jg and Jk in the A, and Jg is in front of Jk which
is at the (v + 1)th position, S1 = {J1 , · · · , Jg , Jk , · · · , Jn }. Suppose that the starting
time of J1 is t0 , d = C[h] , 1 ≤ v < h ≤ n. Z 1 is the objective function of S1 . When
Jg and Jk are swapped, the sequence of jobs is S2 = {J1 , · · · , Jk , Jg , · · · , Jn }. Z 2 is
the objective function of S2 .

v−1

Z 1 − Z 2 = −αt0 (bg − bk ) (1 + b[ j] + r ).
j=1

If Z 1 ≤ Z 2 , bg ≥ bk , i.e. the jobs of set A are arranged in descending order of bi . 



Lemma 4.4 In the optimal scheduling, the jobs of set B are arranged in ascending
order of bi .

Proof There are two adjacent jobs Jg and Jk in the B, and Jg is in front of Jk which
is at the (v + 1)th position, S1 = {J1 , · · · , Jg , Jk , · · · , Jn }. Suppose that the starting
time of J1 is t0 , d = C[h] , h + 1 ≤ v < n. Z 1 is the objective function of S1 . When
Jg and Jk are swapped, the sequence of jobs is S2 = {J1 , · · · , Jk , Jg , · · · , Jn }. Z 2 is
the objective function of S2 .

v−1

Z 1 − Z 2 = βt0 (bg − bk ) (1 + b[ j] + r ).
j=1

If Z 1 ≤ Z 2 , bg ≤ bk , i.e. the jobs of set B are arranged in ascending order of bi . 



123
The due date assignment scheduling... 2179

There are two jobs Jg and Jk in the sequence S1 , and Jg is at the uth position, Jk
is at the vth position, S1 = {J1 , · · · , Jg , · · · , Jk , · · · , Jn }. Suppose that the starting
time of J1 is t0 , bg < bk , d = C[h] , 1 ≤ u ≤ h, h + 1 ≤ v ≤ n. Z 1 is the
objective function of S1 . When Jg and Jk are swapped, the sequence of jobs is S2 =
{J1 , · · · , Jk , · · · , Jg , · · · , Jn }. Z 2 is the objective function of S2 .


u−1 
h−1 
j
Z 1 − Z 2 = t0 (bg − bk ) (1 + b[ j] + r ){−α − α (1 + b[m] + r )
j=1 j=u+1 m=u+1
v−1
 
j
+β (1 + b[m] + r )
j=h+1 m=u+1


h
+ [(h − 1)α + nθ − (n − h)β] (1 + b[ j] + r )}.
j=u+1

Define


h−1 
j v−1
 
j
Suv = −α − α (1 + b[m] + r ) + β (1 + b[m] + r )
j=u+1 m=u+1 j=h+1 m=u+1


h
+[(h − 1)α + nθ − (n − h)β] (1 + b[ j] + r ).
j=u+1

If Suv ≥ 0, Jg should be at the uth position; otherwise, Jg should be at the vth position.
The algorithm is summarized as follows:


n
Algorithm 1 1| p j = b j t, q psd , C O N | (α E j + βT j + θ d)
j=1
Require: t0 , α, β, θ , b j , r
Ensure: The optimal sequence, d
1: First step: All jobs are sorted by increasing deterioration rate, i.e. b[1] ≤ · · · ≤ b[n] .
2: Second step: Determine the optimal position of h =  n(β−θ)
α+β , d = C [h] .
3: Last step: Determine the optimal sequence by Suv .

n
Theorem 4.1 For the problem 1| p j = b j t, q psd , C O N | j=1 (α E j + βT j + θ d), the
complexity of the algorithm is O(nlogn).

Proof The first step requires O(nlogn) time. The second step is completed in con-
stant time. The third step requires O(n) time. So the complexity of the algorithm is
O(nlogn). 


Example 1 There are 4 jobs processed in sequence on a single machine. t0 = 1,

r = 0.2, b1 = 0.3, b2 = 0.2, b3 = 0.4, b4 = 0.1, α = 2, β = 5, θ = 1.

123
2180 J. Qian, H. Han


For the problem 1| p j = b j t, q psd , C O N | nj=1 (α E j + βT j + θ d), the solution
process is given by Algorithm 1.
Step 1 Because b4 ≤ b2 ≤ b1 ≤ b3 , J4 → J2 → J1 → J3 .
Step 2 Calculate the values h =  n(β−θ)
α+β  = 3. d = C [3] .
Step 3 (a) When u = 3 and v = 4, S[3][4] = 3 > 0, J4 is determined at the third
position;
(b) When u = 2 and v = 4, S[2][4] = 1.9 > 0, J2 is determined at the second
position;
(c) When u = 1 and v = 4, S[1][4] = 0.66 > 0, J1 is determined at the first position.
Therefore, the optimal sequence is J1 → J2 → J4 → J3 .

n
5 The problem 1|pj = bj t, qpsd , SLK | j=1 (˛Ej + ˇTj + q)

Lemma 5.1 For any job sequence, q of the optimal scheduling is the (1 + r ) times
completion time of some job or t0 (1 + r ).

Proof Suppose that q of the optimal scheduling isn’t the (1 + r ) times completion
time of some job or t0 (1 + r ), i.e. (1 + r )C[h−1] < q < (1 + r )C[h] , 1 ≤ h ≤ n,
C[0] = t0 . The objective function


n 
h
Z= (α E [ j] + βT[ j] + θq) = −α (1 + r )C[ j−1]
j=1 j=1

n
+β C[ j−1] + [hα + nθ − (n − h)β]q.
j=h+1

(a) When q moves left to (1 + r )C[h−1] , q − (1 + r )C[h−1] = x > 0. The objective

function


h−1 
n
Z 1 = −α (1+r )C[ j−1] +β C[ j−1] +[(h−1)α+nθ −(n−h)β](1+r )C[h−1] ,
j=1 j=h+1

Z − Z 1 = [hα + nθ − (n − h)β]x.
(b) When q moves right to (1 + r )C[h] , (1 + r )C[h] − q = y > 0. The objective
function


h 
n
Z 2 = −α C[ j−1] + β C[ j−1] + [hα + nθ − (n − h − 1)β](1 + r )C[h] ,
j=1 j=h+2

Z − Z 2 = −[hα + nθ − (n − h)β]y.
When hα + nθ − (n − h)β ≤ 0, Z ≥ Z 2 ; otherwise, Z > Z 1 . So q of the optimal
scheduling is the (1 + r ) times completion time of some job or t0 (1 + r ). 


123
The due date assignment scheduling... 2181

Lemma 5.2 For any job sequence, q ofthe optimal

 scheduling is the (1 + r ) times
n(β−θ)
completion time of the job C[h−1] , h = α+β .

Proof When q = (1 + r )C[h−1] , the objective function


h−1 
n
Z = −α (1 + r )C[ j−1] + β (1 + r )C[ j−1]
j=1 j=h+1
+ [(h − 1)α + nθ − (n − h)β](1 + r )C[h−1] .

(a) When q moves left to (1 + r )C[h−2] , the objective function


h−2 
n
Z 1 = −α (1 + r )C[ j−1] + β (1 + r )C[ j−1]
j=1 j=h
+ [(h − 2)α + nθ − (n − h + 1)β](1 + r )C[h−2] .

Z − Z 1 = [(h − 1)α + nθ − (n − h + 1)β](1 + r )(C[h−1] − C[h−2] ) ≤ 0,

n(β − θ )
h≤ + 1.
α+β
(b) When q moves right to (1 + r )C[h] , the objective function


h 
n
Z 2 = −α (1+r )C[ j−1] +β (1+r )C[ j−1] +[hα+nθ −(n−h−1)β](1+r )C[h] .
j=1 j=h+2

Z − Z 2 = [hα + nθ − (n − h)β](1 + r )(C[h−1] − C[h] ) ≤ 0,

n(β − θ )
h≥ .
α+β
n(β−θ) n(β−θ)
So the best position h satisfies α+β ≤h≤ α+β + 1. 

For the convenience of proof, we define two sets: C = {J j ∈ π |C j ≤ q}, D = {J j ∈
π |C j > q}, q = (1 + r )C[h−1] , π is the job sequence.
Lemma 5.3 In the optimal scheduling, the jobs of set C are arranged in descending
order of bi .
Proof There are two adjacent jobs Jg and Jk in the C, and Jg is in front of Jk which
is at the (v + 1)th position, S1 = {J1 , · · · , Jg , Jk , · · · , Jn }. Suppose that the starting
time of J1 is t0 , q = (1 + r )C[h−1] , 1 ≤ v ≤ h − 2. Z 1 is the objective function of S1 .
When Jg and Jk are swapped, the sequence of jobs is S2 = {J1 , · · · , Jk , Jg , · · · , Jn }.
Z 2 is the objective function of S2 .

v−1

Z 1 − Z 2 = −αt0 (1 + r )(bg − bk ) (1 + b[ j] + r ).
j=1

123
2182 J. Qian, H. Han

If Z 1 ≤ Z 2 , bg ≥ bk , i.e. the jobs of set C are arranged in descending order of bi . 



Lemma 5.4 In the optimal scheduling, the jobs of set D are arranged in ascending
order of bi .

Proof There are two adjacent jobs Jg and Jk in the D, and Jg is in front of Jk which
is at the (v + 1)th position, S1 = {J1 , · · · , Jg , Jk , · · · , Jn }. Suppose that the starting
time of J1 is t0 , q = (1 + r )C[h−1] , h ≤ v < n. Z 1 is the objective function of S1 .
When Jg and Jk are swapped, the sequence of jobs is S2 = {J1 , · · · , Jk , Jg , · · · , Jn }.
Z 2 is the objective function of S2 .

v−1

Z 1 − Z 2 = βt0 (1 + r )(bg − bk ) (1 + b[ j] + r ).
j=1

If Z 1 ≤ Z 2 , bg ≤ bk , i.e. the jobs of set D are arranged in ascending order of bi . 



There are two jobs Jg and Jk in the sequence S1 , and Jg is at the uth position, Jk
is at the vth position, S1 = {J1 , · · · , Jg , · · · , Jk , · · · , Jn }. Suppose that the starting
time of J1 is t0 , bg < bk , q = (1 + r )C[h−1] , 1 ≤ u ≤ h − 1, h ≤ v ≤ n. Z 1 is
the objective function of S1 . When Jg and Jk are swapped, the sequence of jobs is
S2 = {J1 , · · · , Jk , · · · , Jg , · · · , Jn }. Z 2 is the objective function of S2 .


u−1 
h−2 
j
Z 1 − Z 2 = t0 (1 + r )(bg − bk ) (1 + b[ j] + r ){−α − α (1 + b[m] + r )
j=1 j=u+1 m=u+1
v−1 
 j
+β (1 + b[m] + r )
j=h m=u+1


h−1
+ [(h − 1)α + nθ − (n − h)β] (1 + b[ j] + r )}.
j=u+1

Define


h−2 
j v−1 
 j
Suv = −α − α (1 + b[m] + r ) + β (1 + b[m] + r )
j=u+1 m=u+1 j=h m=u+1


h−1
+[(h − 1)α + nθ − (n − h)β] (1 + b[ j] + r ).
j=u+1

If Suv ≥ 0, Jg should be at the uth position; otherwise, Jg should be at the vth position.
The algorithm is summarized as follows:

123
The due date assignment scheduling... 2183


n
Algorithm 2 1| p j = b j t, q psd , S L K | (α E j + βT j + θq)
j=1
Require: t0 , α, β, θ , b j , r
Ensure: The optimal sequence, q
1: First step : All jobs are sorted by increasing deterioration rate, i.e. b[1] ≤ · · · ≤ b[n] .
2: Second step: Determine the optimal position of h =  n(β−θ)
α+β , q = (1 + r )C [h−1] .
3: Last step: Determine the optimal sequence by Suv .

n
Theorem 5.1 For the problem 1| p j = b j t, q psd , S L K | j=1 (α E j + βT j + θq), the
complexity of the algorithm is O(nlogn).
Proof The first step requires O(nlogn) time. The second step is completed in con-
stant time. The third step requires O(n) time. So the complexity of the algorithm is
O(nlogn). 



n
6 The problem 1|pj = bj t, qpsd , DIF| (˛Ej + ˇTj + dj )
j=1

Lemma 6.1 In the optimal scheduling, if θ ≥ β, the due date d j of job J j is equal to
0; otherwise, d j is equal to the completion time of job J j .
Proof The objective function


n 
n
Z= Zj = [α max{0, d j − C j } + β max{0, C j − d j } + θ d j ],
j=1 j=1

Z j = α max{0, d j − C j } + β max{0, C j − d j } + θ d j .
(a) When C j ≥ d j ,
Z j = βC j + (θ − β)d j .
If θ ≥ β, d j is equal to 0, Z j = βC j ; if θ < β, d j is equal to C j , Z j = θC j .
(b) When C j ≤ d j ,
Z j = (α + θ )d j − αC j .
So d j is equal to C j , Z j = θC j . 

Lemma 6.2 In the optimal scheduling, the jobs are sequenced in increasing order of
deterioration rate.
Proof We consider the job sequence S1 = {J1 , · · · , Jg , Jk , · · · , Jn }. There are two
adjacent jobs Jg and Jk . Jg is at the vth position in the S1 , Jk is at the (v +1)th position
in the S1 , 1 ≤ v < n. Z 1 is the objective function of S1 . When Jg and Jk are swapped,
the sequence of jobs is S2 = {J1 , · · · , Jk , Jg , · · · , Jn }. Z 2 is the objective function
of S2 .
v−1

Z 1 − Z 2 = min{θ, β}t0 (bg − bk ) (1 + b[ j] + r ).
j=1

123
2184 J. Qian, H. Han

So the jobs are sequenced in increasing order of deterioration rate bi in the optimal
scheduling. 

The algorithm is summarized as follows:


n
Algorithm 3 1| p j = b j t, q psd , D I F| (α E j + βT j + θ d j )
j=1
Require: t0 , α, β, θ , b j , r
Ensure: The optimal sequence, d j
1: First step: The optimal sequence is sequenced by increasing deterioration rate, i.e. b[1] ≤ · · · ≤ b[n] .
2: Last step: Determine the due date d j by β and θ .


n
Theorem 6.1 For the problem 1| p j = b j t, q psd , D I F| (α E j + βT j + θq), the
j=1
complexity of the algorithm is O(nlogn).
Proof The first step requires O(nlogn) time. The second step requires O(n) time. So
the complexity of the algorithm is O(nlogn). 


7 Conclusion

This paper considers the single machine scheduling problem with the CON/SLK/DIF
due date, delivery time and deteriorating jobs. The goal is to minimize the total costs
that contain the earliness, tardiness and due date. Under different conditions, we give
the optimal properties for the problems. We propose the polynomial time algorithms
to obtain the optimal sequence and due date, and the complexity of the algorithms are
O(nlogn). In the future, the multi-machine environment can be considered to expand
the research.

Declarations

Conflict of interest This is the first submission of this manuscript and no parts of this manuscript are being
considered for publication elsewhere. All authors have approved this manuscript. No author has financial
or other contractual agreements that might cause conflicts of interest. We declare that we have no financial
and personal relationships with other people or organizations that can inappropriately influence our work,
there is no professional or other personal interest of any nature or kind in any product, service or company.

References
1. Jackson, J.R.: Scheduling a Production Line to Minimize Maximum Tardiness. Management Science
Research Project, UCLA (1955)
2. Wang, X.Y., Wang, M.Z.: Single machine common flow allowance scheduling with a rate-modifying
activity. Comput. Indus. Eng. 59(4), 898–902 (2010)
3. Li, S., Ng, C.T., Yuan, J.: Scheduling deteriorating jobs with CON/SLK due date assignment on a
single machine. Int. J. Prod. Econ. 131(2), 747–751 (2011)

123
The due date assignment scheduling... 2185

4. Yin, Y.Q., Cheng, T.C.E., Wu, C.C., Cheng, S.R.: Single-machine batch delivery scheduling and
common due-date assignment with a rate-modifying activity. Int. J. Prod. Res. 52(19), 5583–5596
(2014)
5. Zhao, C., Hsu, C.J., Cheng, S.R., Yin, Y., Wu, C.C.: Due date assignment and single machine scheduling
with deteriorating jobs to minimize the weighted number of tardy jobs. Appl. Math. Comput. 248(1),
503–510 (2014)
6. Wang, D.J., Yin, Y., Cheng, S.R., Cheng, T.C.E., Wu, C.C.: Due date assignment and scheduling on a
single machine with two competing agents. Int. J. Prod. Res. 54(4), 1152–1169 (2016)
7. Li, S.S., Chen, R.X.: Common due date assignment and cumulative deterioration scheduling on a single
machine. Eng. Optim. 49(6), 976–989 (2017)
8. Liu, C., Fan, Y., Zhao, C., Wang, J.: Multiple common due-dates assignment and optimal maintenance
activity scheduling with linear deteriorating jobs. J. Indus. Manag. Optim. 13(2), 713–720 (2017)
9. Karhi, S., Shabtay, D.: Single machine scheduling to minimise resource consumption cost with a bound
on scheduling plus due date assignment penalties. Int. J. Prod. Res. 56(9), 3080–3096 (2018)
10. Kim, H.J., Kim, E.S., Lee, J.H.: Scheduling of step-improving jobs with an identical improving rate,
J. Op. Res. Soc. (2021). https://doi.org/10.1080/01605682.2021.1886616
11. Gupta, J.N.D., Gupta, S.K.: Single facility scheduling with nonlinear processing times. Comput. Indus.
Eng. 14(4), 387–393 (1998)
12. Mosheiov, G.: Scheduling jobs under simple linear deterioration. Comput. Indus. Eng. 21(6), 653–659
(1994)
13. Wang, J.B., Guo, Q.: A due-date assignment problem with learning effect and deteriorating jobs. Appl.
Math. Modell. 34(2), 303–309 (2010)
14. Yang, S.H., Wang, J.B.: Minimizing total weighted completion time in a two-machine flow shop
scheduling under simple linear deterioration. Appl. Math. Comput. 217(9), 4819–4826 (2011)
15. Bai, J., Li, Z.R., Huang, X.: Single-machine group scheduling with general deterioration and learning
effects. Appl. Math. Modell. 36(3), 1267–1274 (2012)
16. Wang, X.R., Wang, J.J.: Single-machine scheduling with convex resource dependent processing times
and deteriorating jobs. Appl. Math. Modell. 37(4), 2388–2393 (2013)
17. Yue, Q., Wang, G.: Scheduling Deteriorating Jobs with Common Due Window Assignment. Indus.
Eng. Manag. 20(6), 42–47 (2015)
18. Azadeh, A., Habibnejad-Ledari, H., Zadeh, S.Abdolhossein, Farahani, M.Hosseinabadi: A single-
machine scheduling problem with learning effect, deterioration and non-monotonic time-dependent
processing times. Int. J. Comput. Int. Manufact. 30(2), 292–304 (2017)
19. Zhang, X., Liu, S.C., Yin, Y., Wu, J.B.C.C.: Single-machine scheduling problems with a learning effect
matrix. Iran. J. Sci. Technol. Trans. A Sci. 42(3), 1327–1335 (2018)
20. Wu, C.C., Bai, D., Chung, I.H., Cheng, S.R., Jhwueng, D.C., Lin, W.C., Said, L.B.: A branch-and-
bound and four metaheuristics for minimising total completion time for a two-stage assembly flowshop
scheduling problem with learning consideration. Eng. Optim. 52(6), 1009–1036 (2020)
21. Liang, X.X., Liu, M.Q., Feng, Y.B., Wang, J.B., Wen, L.S.: Solution algorithms for single machine
resource allocation scheduling with deteriorating jobs and group technology. Eng. Optim. 52(7), 1184–
1197 (2020)
22. Wu, C.C., Zhang, X., Azzouz, A., Shen, W.L., Cheng, S.R., Hsu, P.H., Lin, W.C.: Metaheuristics
for two-stage flow-shop assembly problem with a truncation learning function. Eng. Optim. 53(5),
843–866 (2021)
23. Wu, C.C., Bai, D., Chen, J.H., Lin, W.C., Xing, L., Lin, J.C., Cheng, S.R.: Several variants of sim-
ulated annealing hyper-heuristic for a single-machine scheduling with two-scenario-based dependent
processing times. Swarm Evolut. Comput. 60, 100765 (2021)
24. Koulamas, C., Kyparisis, G.J.: Single-machine scheduling problems with past-sequence-dependent
setup times. Euro. J. Op. Res. 187, 1045–1049 (2008)
25. Liu, M.: Parallel-machine scheduling with past-sequence-dependent delivery times and learning effect.
Appl. Math. Modell. 20(6), 42–47 (2013)
26. Yin, Y., Cheng, T.C.E., Xu, J., Cheng, S.R., Wu, C.C.: Single-machine scheduling with past-sequence-
dependent delivery times and a linear deterioration. J. Indus. Manag. Optim. 9(2), 323–339 (2013)
27. Wu, Y.B., Wang, J.J.: Single-machine scheduling with truncated sum-of-processing-times-based learn-
ing effect including proportional delivery times. Neural Comput. Appl. 27, 937–943 (2016)
28. Maecker, S., Shen, L.: Solving parallel machine problems with delivery times and tardiness objectives.
Annals Op. Res. 285(10), 315–334 (2020)

123
2186 J. Qian, H. Han

29. Mönch, L., Shen, L.: Parallel machine scheduling with the total weighted delivery time performance
measure in distributed manufacturing. Comput. Op. Res. 127, 105126 (2021)
30. Graham, R. L., Lawler, E. L., Lenstra, J. K., Kan, A. R.: Optimization and approximation in determin-
istic sequencing and scheduling: a survey, Annals Dis. Math. 5, 287–326 (1979)

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.

123