Hindawi

Mathematical Problems in Engineering

Volume 2021, Article ID 7092981, 10 pages
https://doi.org/10.1155/2021/7092981

Research Article
Joint Pricing and Inventory Control Decisions for Fashion
Apparel with Considering Fashion Level and Partial Backlogging

Lili Zhou , Qi Chen , and Qi Xu

Glorious Sun School of Business and Management, Donghua University, Shanghai 200051, China

Correspondence should be addressed to Qi Chen; [email protected]

Received 22 July 2020; Revised 8 December 2020; Accepted 26 December 2020; Published 5 January 2021

Academic Editor: Petr H jek

Copyright © 2021 Lili Zhou et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Apparel inventories have been crawling up for decades. The fashion level of clothing drops over time, resulting in a continuous
decline in market value, a rapid reduction in demand, and unsold inventory. Poor inventory management weighs down the
apparel industry. Inventory planning decisions are crucial for a fashion apparel retailer. When the good is highly perishable, the
retailer may need to backlog demand in order to market the good at a reasonable price and reduce inventory costs. This study
formulates a finite horizon inventory model for fashion apparel under deteriorating fashion level and partial backlogging
conditions. This model jointly optimizes the selling price, length of inventory-holding period, and length of shortage period by
maximizing the average profit. The conditions that characterize the optimal solution are established and an algorithm is proposed
to search for the optimal solutions. A numerical example is provided to illustrate the algorithm and the solution procedure.
Furthermore, sensitivity analysis of the key system parameters is carried out and some interesting managerial insights are
presented. Findings clearly suggest that the presence of shortage has got an affirmative effect on ordering policy of the fashion
apparel retailer.

1. Introduction implemented fast fashion strategies [2]. Fashion clothing

sales are influenced not only by price but also by fashion
With the rapid development of the world economy and the level, which depends on elements such as style, color, and
advent of the era of globalization, people’s lifestyles have material. Fashion products differ significantly from other
undergone great changes. In addition to comfort, they are types of commercial products. In general, the higher the
also pursuing the individuality and fashion sense of clothing. fashion level of clothing products, the higher their market
According to the statistics from American Fashion [1], in value. As time goes on, the fashion level of the clothing
recent years, the global apparel market has reached 3 trillion drops, resulting in a continuous decline in market value;
US dollars, accounting for about 2% of the global gross market demand may thus fall off rapidly and result in an
domestic product (GDP). Taking the American sports and accumulation of unsold inventory. In other words, fashion
leisure and lifestyle brand Nike as an example, the company apparel decreases in value over the selling season. The latest
is the world’s second largest fashion brand. The company’s data show that excess inventory is a major problem in the
2017 revenue reached US $344.6 billion, and its market value fashion industry. For example, H&M announced at the end
was close to US $105 billion. Fast fashion is an industrial of March 2018 that the total value of its accumulated unsold
practice that is widely applied in fashion retailing. The apparel was US $4 billion. In 2016, the total inventory of 79
central idea is to offer a continuous stream of new mer- quoted textile and apparel companies in China amounted to
chandise to the market which reflects the latest fashion US $81 billion [3]. Inventory planning decisions are crucial
trends and helps capture the hottest designs in the current for supply chain managers in the fashion industry. Having
market. For example, famous international fashion retail short product lifecycle and highly volatile demand, fashion
brands such as H&M, Top Shop, and Zara have all items require extra attention for the initial ordering before
2 Mathematical Problems in Engineering

selling season starts, as well as the subsequent replenishment arise: (1) How does the fashion level of apparel affect product
decisions during the selling season [4, 5]. Bowers and demand? (2) How can we determine the optimal price, the
Agarwal [6] proposed a dynamic hierarchical model for optimal length of inventory-holding period, and the optimal
apparel production planning and scheduling in which they length of shortage period to maximize the apparel retailer’s
considered all processes from daily scheduling and short- average profit during the selling period? (3) How the inventory
term production planning to long-term decisions to better control parameters, such as the deteriorating rate of the fashion
match the demand with a minimal total cost. Fisher et al. [7] level, the backlogging resistance, the shortage cost, and in-
studied replenishment order quantity to minimize the cost ventory-holding cost, influence apparel retailer’s performance?
of lost sales and obsolete inventory in a two-stage model. In our inventory model, the shortage is allowed, and the partial
They identified the optimal reorder time and quantified the backlogging rate is sensitive to the waiting time. Firstly, we
benefit of lead time reduction. Raman and Kim [8] quan- prove that, for any given replenishment cycle, the average
tified the impact of inventory-holding cost in the fashion profit is strictly concave in selling price and, for any given
supply chain based on data from Northco, a school uniform selling price, the average profit function is strictly pseudo-
manufacturer in Northeastern USA. However, the impact of concave in replenishment cycle; that is, there exists a unique
the fashion level of clothing products on pricing and re- and optimal solution to the problem. Next, an algorithm is
plenishment strategies of apparel retailer has been largely proposed to obtain the optimal solution of the problem. Fi-
neglected in prior studies [6–8]. Due to the time-varying nally, the sensitivity analyses of the parameters are conducted
characteristics of fashion clothing, implementing a joint to demonstrate the effectiveness of the proposed model. The
pricing and replenishment strategy has obvious practical research results show that when the purchasing cost, the
significance for managing inventory replenishment and shortage cost, and the holding of fashion apparel increase, the
improving net profits. Chen et al. [3] analyzed the joint average profit of system will decrease. The optimal average
optimal dynamic pricing strategy and replenishment cycle profit and order quantity are highly sensitive to the deterio-
for fashion apparel considering the effect of fashion level. ration rate of fashion level and the backordering resistance.
The decay of fashion over time leads to a decrease in clothing The rest of this paper is organized as follows. In Section
sales demand, which makes the storage time of products 2, we briefly review the related literature. In Section 3, we
longer and ultimately leads to an increase in inventory costs. outline the notations and assumptions used to address the
In reality, clothing retailers can reduce inventory costs by problem. Section 4 contributes to the formulation of the
delaying in meeting demand of consumers, that is, allowing model. We obtain theoretical results and provide an iterative
the products to be temporarily out of stock or partly out of algorithm in Section 5. Numerical examples, sensitivity
stock to delay meeting consumers’ needs. When it is eco- analysis, and some managerial insights are provided in
nomic to backlog demand, the reseller can plan for periods Section 6. Finally, Section 7 offers a conclusion and discusses
of shortage during which demand can be partially back- the implications for management.
ordered [9]. When the good is highly perishable, the retailer
may need to backlog demand in order to market the good at 2. Literature Review
a reasonable price. Apart from the direct effects of fashion
level and demand rate on changes in inventory levels, ap- To date, the literature on fashion inventory management has
parel inventory will inevitably suffer from a shortage. In focused on quick response strategies. In the fashion industry,
some inventory systems, such as fashionable items, the more and more companies, such as Uniqlo, Zara, Topshop,
length of the waiting time for next replenishment would and Primark, are adopting such strategies [11, 12]. Quick
determine whether the backlogging will be accepted or not. response strategies were first proposed in the 1980s in the US
Therefore, the backlogging rate is variable and is dependent fashion industry [13]. Fashion retailers can postpone order
on the waiting time for the next replenishment [10]. To placement until close to the beginning of the selling season to
enhance the exploratory power and the connection between obtain more accurate market information about the sales of
research and practice, we develop our theoretical framework particular types of fashion items (i.e., style trends). This allows
learning from the study by Chen et al. [3] on inventory them to adjust their initial inventory forecast, so that supply
management for fashion apparel. Chen et al. [3] used a can better match demand. Other related studies include that
dynamic pricing model to examine the effect of fashion level by Aviv et al. [14], where they studied the potential benefits of
on firm performance. We draw upon inventory planning response pricing for seasonal fashion merchandise sellers.
decisions theory as a response to the call for an increased Cachon and Swinney [15] discuss the quick response medi-
focus on the partial backlogging and contribute to the ated fast fashion program in the apparel industry. Chow et al.
broader application of inventory planning decisions theory [16] study the impacts of imposing a minimum order quantity
in inventory management for fashion apparel. In addition, on a two-echelon supply chain implementing the quick re-
this study also contributes to the deterioration items liter- sponse strategies. Choi [17] studied a sourcing decision with
ature by simultaneously considering the fashion level and the quick response strategy under an environmental sus-
partial backlogging. tainability practice. Chan et al. [18] determined the number of
Therefore, this paper aims to focus on the pricing and observations to take under quick response for apparel in-
replenishment strategy problems for the fashion apparel re- ventory control. Choi [19] constructed a formal analytical
tailer with considering fashion level and partial backlogging. fashion supply chain model based on the Bayesian infor-
Based on this research background, the following questions mation updating to examine the effects of quick response.
Mathematical Problems in Engineering 3

Cosgun et al. [20] addressed the simultaneous determination model to examine the effect of fashion level on firm per-
of markdown prices and optimal initial inventory levels under formance. From the aspect of modeling, the most closely
the cross-price effects in a random demand setting for related studies to this paper are those by Tsao [29] and Chen
multiproduct groups. All of these studies proposed using et al. [3]. However, unlike these studies, we incorporate the
quick response strategies to deal with volatile market demand fashion level and shortage into the optimization model. This
by reducing order lead time to more accurately determine the paper derives a pricing and replenishment policy for fashion
appropriate quantity to order. apparel and adopts a price- and time-dependent demand
Inventory planning decisions are crucial for fashion function to model the finite time horizon inventory. The
apparel. As such apparel has a short product lifecycle and is major objective of this policy is to simultaneously determine
associated with highly volatile demand, initial orders before the optimal price, optimal length of inventory-holding pe-
the selling season begins require careful attention [4, 5], and riod, and optimal length of shortage period to maximize
the same applies for replenishment decisions during the average profit. We also establish the conditions that char-
selling season. Due to the complexity of fashion inventory acterize the optimal solutions and propose an algorithm to
management, very few studies have focused on fashion search for optimal solutions. To be specific, a systematic
inventory management from a pricing and replenishment comparison between this paper and other related papers is
policy perspective. Deterioration items exist commonly in given in Table 1 to show the novelty and advantage of this
the real business environment. As stated in previous liter- paper.
ature [21, 22], deterioration refers to spoilage, decay, pil-
ferage, decay, evaporation, obsolescence, or damage such 3. The Model
that the items cannot maintain the original value or utility.
Most of the physical goods undergo decay or deterioration 3.1. Problem Formulation. First, we describe the problem as
over time, such as medicine, volatile liquids, fashion items, follows. We model an apparel retailer that sells fashionable
and others [23]. For the property of the deterioration items, clothing products. The demand for fashionable clothing is
the research on pricing and replenishment strategy for the determined by the selling price p and fashion level of the
fashion apparel retailer can be referred to the research on clothing ω(t) during the selling period. The replenishment
inventory and pricing decisions of perishable products problem for a single fashion item is considered. The in-
[24–26]. For example, Wu et al. [24] formulated an in- ventory system evolves as follows: Q units of items arrive at
ventory model with deteriorating items and price-sensitive the inventory system at the beginning of each cycle. The
demand and obtained the retailer’s optimal selling price and inventory level I(t) decreases due to consumers’ demand
the length of replenishment cycle. Avinadav et al. [25] used a during the replenishment cycle. During the interval
price- and time-dependent demand function to develop a 0 ≤ t ≤ t1 , the inventory level decreases due to demand. A
mathematical model that calculates the optimal price and shortage occurs due to demand and partial backlogging
replenishment period of perishable items. Herbon et al. [26] during the time interval t1 ≤ t ≤ t1 + t2 . The length of re-
formulated an inventory model with deteriorating items and plenishment cycle T equals the length of inventory-holding
price-sensitive demand and obtained the retailer’s optimal period t1 plus the length of shortage period t2 . The back-
dynamic selling price and the length of replenishment cycle. logged demand during the shortage period is S. Thus, the
In the above-mentioned inventory models, the shortage of retailer’s ordering quantity for one replenishment cycle is
products has been neglected. The shortage setting should be Q � I(0) + S. The selling price, length of inventory-holding
considered to face the various practical requirements. period, and length of shortage period are decision variables.
Nevertheless, few research efforts have explicitly con- The major objective of this policy is to simultaneously de-
sidered the inventory decisions of fashion apparel from the termine the selling price, the length of inventory-holding
perspective of pricing and replenishment cycles. Fisher et al. period, and the length of shortage period to maximize av-
[27] studied how replenishment order quantity may mini- erage profit.
mize the cost of lost sales, backorders, and obsolete in-
ventory in a two-stage model. Kogan and Spiegel [28] 3.2. Assumptions. The mathematical model in this paper is
developed a model with an initial inventory level and developed on the basis of the following assumptions.
addressed the problem of maximizing profits from the
production, storage, and sale of fashion goods. In their
Assumption 1. The demand function is nonnegative and
proposed inventory model, the selling price remains un-
explicitly depends separately on both the price and fashion
changed throughout the sales period. Such a price strategy
level. In practice, the more fashionable a garment, the higher
ignores the fashion level of the apparel, which declines over
its market value, increasing consumers’ incentive to buy it
time. Tsao [29] developed an analytical model of inventory
[3]. That is, the higher the fashion level is, the larger the
control for deteriorating goods and fashion goods under
demand will be, according to the following formula:
trade credit and partial backlogging conditions. In that
study, two-phase pricing and inventory decisions were d(p) + ω(t), 0 ≤ t ≤ t1 ,
considered and the optimal replenishment strategy was D(p, t) � 􏼨 (1)
d(p), t1 ≤ t ≤ t1 + t2 .
identified. The above-mentioned works of literature have not
considered the influence of fashion level on pricing and Without loss of generality, let the function d(p) be
inventory decisions. Chen et al. [3] used a dynamic pricing linear; that is, d(p) � a − bp. a is the fixed term of demand
4 Mathematical Problems in Engineering

Table 1: A comparison of the present work with related previous works.

References Quick response Inventory planning decisions Fashion level Shortage
[14–19] Yes No No No
[24–26] No Yes No No
[3] No Yes Yes No
[29] No Yes No Yes
Our paper No Yes Yes Yes

determined by the market size and b represents the sensi- With the condition I(t1 ) � 0, solving formula (4) yields
t
tivity of demand with respect to price. The demand function I1 (t) � 􏽒t1 (d(p) + ω(u))du. Therefore, the initial inventory
t
of paper is based on several previous studies of perishable level is then given by I0 � I1 (0) � 􏽒01 (d(p) + ω(u))du.
products. In those studies, the demand function was usually During the second interval t1 ≤ t ≤ t1 + t2 , shortage occurred,
assumed to have a linear form. This assumption has been and the demand is partially backlogged according to the
widely adopted in the perishable inventory management fraction B(T − t). That is, the inventory level at time t is
literature, such as Li et al. [22], Zhang et al. [23], and Herbon governed by the following differential equation:
et al. [26]. dI2 (t)
� − d(p)B t1 + t2 − t􏼁. (5)
dt
Assumption 2. The fashion level at time t is given by
With the condition I(t1 ) � 0, solving (5) yields
t
ω(t) � ω0 e− ηt . (2) I2 (t) � − d(p) 􏽒t2 +t − t B(τ)dτ. If we put t � T into the above
1 2
formula, the maximum amount of demand backlogging will
Equation (2) is motivated by the model presented in be obtained as follows: S � d(p)M(t2 ), where
Chen et al.’s work [3] and is proposed based on the following t
M(t2 ) � 􏽒02 B(τ)dτ. Order quantity per cycle is the total of S
consideration. The fashion level function is strictly de- and I0 ; that is,
creasing with time t; dω/dt < 0. Therefore, we can deduce
t1
that the demand is decreasing with time t, so the assumption Q � 􏽚 (d(p) + ω(u))du + d(p)M t2 􏼁. (6)
of the demand function in (1) is reasonable. 0

Thus, the inventory level during the replenishment cycle

Assumption 3. Shortages are allowed. We adopt the notation is described as follows:
used in Abad [9] where the unsatisfied demand is back- t1
logged, and the fraction of shortage back-ordered is ⎪
⎧
⎪
⎪
⎪ 􏽚 (d(p) + ω(u))du, 0 ≤ t ≤ t1 ,
⎨ t
B(τ) � k0 e − k1 τ
. (3) I(t) � ⎪ t2
(7)
⎪
⎪
⎪
⎩ − d(p) 􏽚 B(τ)dτ, t1 ≤ t ≤ t1 + t2 .
where 0 < k0 < 1 is the backlogging intensity, τ > 0 is the t1 +t2 − t

waiting time up to the next replenishment, and k1 > 0 is the Now, we can obtain inventory costs and sales revenue
backlogging resistance. Note that B(τ) proportions of the per cycle which consist of the following five elements: fixed
consumers who arrive during the shortage period are order cost A, sales revenue SR � pQ, purchase cost PC � cQ,
willing to wait for the next replenishment resulting in the λ
inventory cost HC � h 􏽒0 I1 (t)dt, and shortage cost
backlogged demand, whereas the remaining 1 − B(τ) t2
SC � sd(p) 􏽒0 τB(τ)dτ, where p is the selling price per unit,
proportions of the consumers are lost, resulting in lost sales c is the purchasing cost per unit, h is inventory-holding cost
[22]. From formula (3), it can be seen that zB(τ)/zτ < 0 and per unit, and s is the shortage cost per unit.
z2 B(τ)/zτ 2 > 0, and the partial backlogging decreases Therefore, the total profit (denoted by Π(t1 , t2 , p))
rapidly with the increase of waiting time, indicating that during the time interval [0, T] is given by
customers are becoming impatient, which is consistent
with the reality. Π t1 , t2 , p􏼁 � SR − PC − HC − SC − OC − A
t1

4. Model Formulation � (p − c)􏼨􏽚 (d(p) + ω(u))du + d(p)M t2 􏼁􏼩

0
t1 t1
From Figure 1, we can see that the inventory level I(t)
− h 􏽚 􏽚 (d(p) + ω(u))dudt
decreases due to consumers’ demand during the time in- 0 t
terval 0 ≤ t ≤ t1 . The differential equation that represents t2
inventory status is given by − s d(p) 􏽚 τB(τ)dτ.
0
dI1 (t) (8)
� − d(p) − ω(t). (4)
dt
Mathematical Problems in Engineering 5

I (t) The length of shortage period t2 falls in the range of

(0, (p − c)/s]. For any given selling price p, the objective
function (9) is equivalent to the following programming
Q I1 (t) problem:
max π t1 , t2 |p􏼁,
t2
⎪
⎧ t1 ≥ 0,
⎪
⎪ (13)
⎨
0 s.t. ⎪
⎪p− c
⎪
t1 T ⎩ ≥ 0.
I2 (t) s − t2
S Shortage
Note that formula (13) represents fractional programs. A
fractional program is defined as the nonlinear program sup
Lost sales
Φ(x)/Γ(x), x ∈ U � {x ∈ X: h(x) ≤ 0}, where Γ(x) > 0. If
Figure 1: The apparel inventory level. Φ(x) > 0 is concave, Γ(x) and h(x) are convex; then the
problem is a concave fractional program. For a differentiable
and strictly concave fractional program, there is at most one
The average profit functions are given by maximum, and a solution to the Kuhn-Tucker conditions is
a maximum. We refer readers to Li et al.’s work [22] for the
Π t1 , t2 , p􏼁
π t1 , t2 , p􏼁 � . (9) details on the definition and properties of fractional
t1 + t2 program. □

Proposition 2. For any given p, the average profit π(t1 , t2 |p)

5. Results and Algorithm is strictly pseudoconcave in (t1 , t2 ).
The retailer’s objective is to maximize π(t1 , t2 , p) by jointly
determining (t1 , t2 , p). In this section, we will explore the Proof. From equation (9), it is easy to get z2 Π/zt21 < 0,
mathematical properties of the model to assist in proposing z2 Π/zt1 zt2 � 0, and z2 Π/zt22 � d(p)[(p − c − st2 )B′ (t2 )
an algorithm to search for the optimal solution. − sB(t2 )]. Since 0 ≤ B(t2 ) ≤ 1 and B′ (t2 ) ≤ 0, we get
The necessary conditions for the average profit (9) to be z2 Π/zt22 < 0. Let H denote Hessian matrix of Π; then we have
maximized are H � (z2 Π/zt21 )(z2 Π/zt22 ) − (z2 Π/zt1 zt2 )(z2 Π/zt1 zt2 ) > 0.
Therefore, H is negative definite, and Π(t1 , t2 , p) is strictly
zπ t1 , t2 , p􏼁 1 zΠ concave in (t1 , t2 ). Also, the denominator t1 + t2 is a linear
� � 0, (10)
zp t1 + t2 􏼁 zp function. Hence, π(t1 , t2 |p) is strictly pseudoconcave in
(t1 , t2 ).
zπ t1 , t2 , p􏼁 zΠ/zt1 􏼁 t1 + t2 􏼁 − Π Next, to simplify the expression of the model, let
� 2 � 0, (11)
zt1 t1 + t2 􏼁 Δ � (p − c)ω(0)g(0) + A + d(p)
g(0) (14)
zπ t1 , t2 , p􏼁 zΠ/zt2 􏼁 t1 + t2 􏼁 − Π 􏼨(p − c)(g(0) − M(g(0))) + s 􏽚 τB(τ)dτ 􏼩,
� 2 � 0. (12) 0
zt2 t1 + t2 􏼁

where g(0) � t2 |G(t1 �0,t2 )�0. We also define the roots of

Proposition 1. For any given t1 and t2 , the average profit zπ(t1 , t2 , p)/zt1 � 0 and zπ(t1 , t2 , p)/zt2 � 0 as t+1 and t+2 ,
π(t1 , t2 , p) is strictly concave in p, and there exists a unique respectively. □
global maximum point p∗ ∈ (c, ∞) that satisfies the first-
order necessary condition zπ/zp � 0. Proposition 3. For any given p ∈ (c, ∞), π(t1 , t2 |p) has a
global maximum point (t∗1 , t∗2 ), which is determined by the
following rules:
Proof. From equation (8), we get z2 Π/zp2 � − 2b
(t1 + M(t2 )) < 0. According to equation (10), we conclude (1) If Δ ≤ 0, then (t∗1 , t∗2 ) � (0, g(0))
that z2 π/zp2 � (1/(t1 + t2 ))(z2 Π/zp2 ) < 0; therefore, π is (2) If Δ ≥ 0, then (t∗1 , t∗2 ) � (t+1 , t+2 )
strictly concave in p. In addition, we have limp⟶+∞
(zπ/zp) < 0 and limp⟶c (zπ/zp) > 0. Therefore, there exists
a unique maximum point p∗ ∈ (c, ∞) to satisfy zπ(t1 , Proof. Let λ1 and λ2 be two Lagrange Multipliers. Kuhn-
t2 , p)/zp � 0. Tucker condition of equation (13) is
6 Mathematical Problems in Engineering

⎪
⎧ zπ t1 , t2 , p􏼁 (b) If Δ ≤ 0, then limλ⟶0+ F(t1 ) < 0. Therefore,
⎪
⎪ + λ1 � 0,
⎪
⎪ zt1 (zπ/zt1 ) < 0, and then λ1 � − (zπ/zt1 ) > 0, which
⎪
⎪
⎪
⎪ means that Case (iii) is feasible and t∗1 � 0. Subse-
⎪
⎪
⎪
⎪ quently, we get t∗2 by substituting t∗1 into equation
⎪
⎪ zπ t1 , t2 , p􏼁zt2 − λ2 � 0,
⎪
⎪ (18). From Proposition 2, the local maximum point
⎪
⎪
⎪
⎪ (t∗1 , t∗2 ) is the global maximum point.
⎨
⎪ λ1 t1 � 0, (15) Not surprisingly, similar to the works of Li et al. [22] and
⎪
⎪
⎪
⎪ Tsao [29], it is very difficult to solve the nonlinear equations
⎪
⎪
⎪
⎪ (p − c) (10)–(12). Based on the preceding propositions, we use a
⎪
⎪
⎪
⎪ λ􏼠 − t1 􏼡 � 0, Gauss-Newton iterative method and propose the following
⎪ 2
⎪ s
⎪
⎪ algorithm to obtain the optimal solution:
⎪
⎪
⎪
⎪
⎩ Step 1: Set i � 1, and initialize the value of p � p(i) > c.
λ1 , λ2 ≥ 0.
Step 2: Calculate Δ(p(i) ). According to Proposition 3,
To solve the above equations, we can consider the three execute one of the following substeps:
following cases:
(1) If Δ(p(i) ) ≥ 0, then t(i) + (i) +
1 � t1 and t2 � t2 .
Case (i). If λ2 ≠ 0, then t2 � (p − c)/s and zπ/zt2 < 0,
which contradicts the constraint of λ2 ≥ 0. Therefore, (2) If Δ(p(i) ) ≤ 0, then t(i) (i)
1 � 0 and t2 � g(0).
this scenario will not occur.
Step 3: Solve zπ(λ(i) , μ(i) , p)/zp � 0 and mark the so-
Case (ii). If λ1 � 0 and λ2 � 0, then zπ(t1 , t2 |p)/zt1 � 0
lution as p � p(i+1) . If the difference between p(i) and
and zπ(t1 , t2 |p)/zt2 � 0.
p(i+1) is small enough, such as |p(i+1) − p(i) | < 10− 6 , the
Case (iii). If λ1 ≠ 0 and λ2 � 0, then t1 � 0, optimal solution is (t∗1 , t∗2 , p∗ ) � (t(i) (i) (i+1)
1 , t2 , p ).
zπ(t1 , t2 |p)/zt2 � 0, and λ1 � − zπ(t1 , t2 |p)/zt1 . Otherwise, set i � i + 1 and go back to Step 2.
Next, we prove that case (ii) and case (iii) are feasible. Step 4: We can obtain Q∗ using (6) and π∗ using (9),
Equations (11) and (12) are equivalent to respectively. □
zΠ
− t1 + t2 􏼁 � Π, (16) 6. Numerical Analysis
zt1
In this section, we conduct numerical analyses to gain
zΠ
− t1 + t2 􏼁 � Π. (17) managerial insights. The basic parameter values are chosen
zt2 mainly based on the work of Chen et al. [3] and satisfy the
From equations (16) and (17), we get assumptions mentioned in the Model Formulation section.
The parameters are designed as follows: market size
zΠ zΠ a � 1000, price sensitivity parameter b � 20, purchasing cost
− � 0. (18) per unit c � 40, inventory-holding cost per unit h � 2,
zt1 zt2
shortage cost per unit s � 3, k0 � 0.9, k1 � 0.01, the dete-
We define G(t1 , t2 ) � (zΠ/zt1 ) − (zΠ/zt2 ). Since riorating rate of the fashion level η � 0.02, initial fashion
zG(t1 , t2 )/zt2 > 0, G(t1 , t2 ) increases as t2 increases. More- level ω0 � 100 , and fixed cost per order A � 400. Then the
over limt2 ⟶ ((p− c)/s)G(t1 , t2 ) � (zΠ/zt1 ) > 0 and limt2 ⟶ 0+ corresponding optimal solutions are given as follows: the
G(t1 , t2 ) � (c − p − ht1 )ω(t1 ) − ht1 d(p) < 0. Thus, there optimal price p∗ � 47.4156, the optimal length of inventory-
exists a unique t2 ∈ (0, (p − c)/s) such that G(t1 , t2 ) � 0. holding period t∗1 � 2.6312, the optimal length of shortage
Therefore, t2 can be regarded as the function of t1 . Let t2 � period t∗2 � 0.7487, the optimal order quantity
g(t1 ) and substitute it into equation (16); then we get Q∗ � 452.6448, and the optimal average profit
(zΠ/zt1 )(t1 + g(t1 )) − Π � 0. Let F(t1 ) � (zΠ/zt1 )(t1 + π∗ (t1 , t2 , p) � 890.806. Next, we examine the impacts of the
g(t1 )) − Π; then we obtain the first derivative of F(t1 ) with related parameters on the optimal solutions and obtain some
respect to t1 : (dF(t1 )/dt1 ) � (z2 Π/zt21 )(t1 + g(t1 )) < 0. In useful insights.
addition, limt1 ⟶ 0+ (zΠ/zt1 ) � (p − c)(d(p) + ω0 ) > 0 and
limt1 ⟶ ∞(zΠ/zt1 ) < 0. Thus, there exists a unique
t1 ∈ [0, +∞) such that (zΠ/zt1 )|t1 �t1 � 0. Therefore, F(t1 ) is 6.1. Impact of the Deteriorating Rate of the Fashion Level (η).
strictly decreasing in t1 ∈ [0, t1 ], and we get The optimal selling price p∗ , optimal length of inventory-
limt1 ⟶ t F(t1 ) � − Π < 0. Moreover, limλ⟶0+ F(t1 ) � Δ. holding period t∗1 , optimal length of shortage period t∗2 , and
optimal average profit π∗ decrease in η (Table 2). Table 2
(a) If Δ ≥ 0, then limt1 ⟶ 0+ F(t1 ) > 0. By Intermediate
shows the effect of deterioration rate η of the fashion level.
Value Theorem, there exists a unique t+1 ∈ [0, t1 ]
The smaller the deterioration rate, the slower the lowering of
satisfying equation (16). Thus, Case (ii) is feasible,
fashion level because the fashion level largely depends on the
and t∗1 � t+1 .
Mathematical Problems in Engineering 7

Table 2: Impact of the deteriorating rate of the fashion level.

η p∗ t∗1 t∗2 π∗
0.02 47.41 2.63 0.7487 890
0.03 47.40 2.61 0.7487 885
0.04 47.39 2.59 0.7486 884
0.05 47.37 2.57 0.7485 883
0.06 47.36 2.56 0.7482 881
0.07 47.35 2.54 0.7478 879
0.08 47.34 2.52 0.7475 877
0.09 47.32 2.50 0.7470 873

elements of style and material, all of which affect the de- profit π∗ decrease in s. The optimal average profit is very
terioration rate. Therefore, the larger the deterioration rate, sensitive to the change of shortage cost. When the
the faster the attenuation of fashion products. The faster shortage cost increases, the clothing retailer should reduce
attenuation of fashion level leads to the shorter sales cycle. the order quantity in order to obtain higher profit. Fur-
Then the market demand will be correspondingly reduced, thermore, the results of Tables 2–4 show that the optimal
resulting in the lower average profit. It can be seen from average profit π∗ increases with the optimal length of
Table 2 that the change of the deterioration rate has a great shortage period t∗2 . It suggests that the presence of
influence on average profit but little influence on optimal shortage has got an affirmative effect on fashion apparel
pricing and replenishment period. The results show that retailer ordering policy.
when the deterioration rate of fashion apparel increases, the
fashion products retailer should pay attention to reducing
the selling price appropriately, reducing the inventory- 6.4. Sensitivity Analysis of Other Parameters (a, b, c, h).
holding period and out of stock period, and reducing the We now study the effects of changes in the values of the
quantity of orders as much as possible, so as to improve the parameters a, b, c, and h on the optimal selling price p∗ ,
average profit. order quantity Q∗ , and average profit π∗ . The sensitivity
analysis is performed by changing each value of the pa-
rameters by − 60%, − 40%, − 20%, 0%, +20%, +40%, and
6.2. Impact of the Backlogging Resistance (k1 ). As can be seen
+60%, taking one parameter at a time and keeping the
from Table 3, the optimal selling price p∗ and optimal length
remaining parameter values unchanged. The computational
of inventory-holding period t∗1 increase in k1 , whereas the
results are shown in Figure 2.
optimal length of shortage period t∗2 and optimal average
The sensitivity analysis shown in Figure 2 indicates the
profit π∗ decrease in k1 . The larger the backlogging re-
following observations.
sistance is, the more reluctant the customer is to wait. In
The optimal selling price p∗ decreases in the values of
this case, the inventory-holding time should be appro-
parameters b, c, and h. Moreover, p∗ is lowly positive
priately extended, the shortage time should be shortened,
sensitive to changes in parameters c and h, whereas p∗ is a
and the total order quantity should be correspondingly
highly positive sensitive to change in c. It is reasonable that
reduced. It can be seen from Table 3 that the change of the
the purchase cost has a great and positive effect upon the
backlogging resistance has a great influence on average
optimal price.
profit but little influence on optimal pricing and replen-
When the values of parameters b, c, and h increase, the
ishment period. Therefore, for customer groups with dif-
optimal order quantity Q∗ will decrease. Q∗ is moderately
ferent backlogging resistance, the corresponding pricing
sensitive to changes in parameters c and h, whereas Q∗ is
and replenishment strategies will also be very different,
highly sensitive to changes in a and b.
which needs attention of clothing retailers.
When the values of parameters b, c, and h increase, the
optimal average profit π∗ decreases. π∗ is moderately
6.3. Impact of the Shortage Cost (s). As can be seen from sensitive to changes in parameters c and h, whereas Q∗ is
Table 4, the optimal selling price p∗ and optimal length of highly sensitive to changes in a and b. These results indicate
inventory-holding period t∗1 increase in s, whereas the that increases in cost have a negative effect on the average
optimal length of shortage period t∗2 and optimal average profit.
8 Mathematical Problems in Engineering

Table 3: Impact of the backlogging resistance.

k1 p∗ t∗1 t∗2 π∗
0.010 47.41 2.61 0.7487 890
0.015 47.89 2.62 0.7295 886
0.020 48.07 2.64 0.7134 881
0.025 48.40 2.65 0.6952 877
0.030 48.91 2.68 0.6790 873

Table 4: Impact of the shortage cost.

∗
s p t∗1 t∗2 π∗
2.0 47.22 2.58 0.8435 901
2.5 47.31 2.61 0.7719 893
3.0 47.41 2.63 0.7487 891
3.5 47.52 2.67 0.7376 886
4.0 47.62 2.68 0.7180 879

150 150

100 100
Variable change (%)

Variable change (%)

50
50

0
0

–50
–50

–100
–100
–50 0 50 –50 0 50
a change (%) b change (%)

Q∗ (% change) Q∗ (% change)
p∗ (% change) p∗ (% change)
π∗ (% change) π∗ (% change)
(a) (b)
100
100

50
50
Variable change (%)

Variable change (%)

0 0

–50 –50

–100 –100
–50 0 50 –50 0 50
c change (%) h change (%)
Q∗ (% change) Q∗ (% change)
p∗ (% change) p∗ (% change)
π∗ (% change) π∗ (% change)
(c) (d)

Figure 2: Effect of changes in values of parameters a, b, c, and h on the model variables.

Mathematical Problems in Engineering 9

7. Conclusions Acknowledgments
The fashion level of clothing drops over time, resulting in a This work was supported by the National Natural Science
continuous decline in market value, a rapid reduction in Foundation of China (71572033 and 71832001).
demand, and unsold inventory. In reality, clothing re-
tailers can reduce inventory costs by delaying in meeting
demand of consumers, that is, allowing the products to be
References
temporarily out of stock or partly out of stock to delay [1] Global fashion industry statistics International apparel,
meeting consumers’ needs. Thus, an inventory model is https://fashionunited.com/global-fashion-industry-statistics.
proposed for deteriorating fashion level apparel under [2] T.-M. Choi, C.-L. Hui, N. Liu, S.-F. Ng, and Y. Yu, “Fast
shortage policy. In the model, the shortage is allowed, and fashion sales forecasting with limited data and time,” Decision
the partial backlogging rate is sensitive to the waiting time. Support Systems, vol. 59, pp. 84–92, 2014.
Firstly, we prove that, for any given replenishment cycle, [3] Q. Chen, Q. Xu, and W. Wang, “Optimal policies for the
the average profit is strictly concave in selling price and, for pricing and replenishment of fashion apparel considering the
any given selling price, the average profit function is effect of fashion level,” Complexity, vol. 2019, Article ID
9253605, 12 pages, 2019.
strictly pseudoconcave in replenishment cycle; that is,
[4] B. Shen, H. L. Chan, P. S. Chow, and K. A. T. Barletta,
there exists a unique and optimal solution to the problem. “Inventory management research for the fashion industry,”
Next, an algorithm is proposed to obtain the optimal price, International Journal of Inventory Research, vol. 3, no. 4,
optimal length of inventory-holding period, and optimal pp. 297–317, 2016.
length of shortage period, so as to maximize the profit. [5] X. Wen, T.-M. Choi, and S.-H. Chung, “Fashion retail supply
Finally, the sensitivity analyses of the parameters are chain management: a review of operational models,” Inter-
conducted to demonstrate the effectiveness of the pro- national Journal of Production Economics, vol. 207, pp. 34–55,
posed model. The research results show that when the 2019.
purchasing cost, the shortage cost, and the holding of [6] M. R. Bowers and A. Agarwal, “Hierarchical production
fashion apparel increase, the average profit of system will planning,” International Journal of Clothing Science and
decrease. The optimal average profit and order quantity are Technology, vol. 5, no. 3/4, pp. 36–43, 1993.
highly sensitive to the deterioration rate of fashion level [7] M. Fisher, K. Rajaram, and A. Raman, “Optimizing inventory
and the backordering resistance. The presence of shortage replenishment of retail fashion products,” Manufacturing &
Service Operations Management, vol. 3, no. 3, pp. 230–241,
has got an affirmative effect on ordering policy of the
2001.
fashion apparel retailer. [8] A. Raman and B. Kim, “Quantifying the impact of inventory
In this paper, we mainly focus on the impact of fashion holding cost and reactive capacity on an apparel manufac-
level and products’ shortage on the operational policy. turerʼs profitability,” Production and Operations Management,
Although we take the pricing decision into account in the vol. 11, no. 3, pp. 358–373, 2002.
model, the attention on the pricing policy is limited, since it [9] P. L. Abad, “Optimal pricing and lot-sizing under conditions
is not affected by fashion level and time. In fact, dynamic of perishability and partial backordering,” Management Sci-
pricing is a powerful tool to manage the demand changing ence, vol. 42, no. 8, pp. 1093–1104, 1996.
over time products. Dynamic pricing is a price discrimi- [10] R. Maihami and I. Nakhai Kamalabadi, “Joint pricing and
nation based on the time of purchase by customers. It can inventory control for non-instantaneous deteriorating items
reduce the fluctuation of demand to a certain extent and with partial backlogging and time and price dependent de-
increase the company’s income. Therefore, the problem of mand,” International Journal of Production Economics,
dynamic pricing strategy and replenishment cycle for vol. 136, no. 1, pp. 116–122, 2012.
[11] Y.-T. Lin and A. Parlaktürk, “Quick response under com-
fashion apparel by considering the effect of fashion level
petition,” Production and Operations Management, vol. 21,
and products’ shortage would be worth exploring in the no. 3, pp. 518–533, 2012.
future. In addition, this paper considers the deterministic [12] T.-M. Choi, “Impacts of retailer’s risk averse behaviors on
situation, thus considering the stochastic situation, such as quick response fashion supply chain systems,” Annals of
stochastic demand, which can be another future research Operations Research, vol. 268, no. 1, pp. 239–257, 2018.
direction. [13] T.-M. Choi and S. Sethi, “Innovative quick response pro-
grams: a review,” International Journal of Production Eco-
nomics, vol. 127, no. 1, pp. 1–12, 2010.
Data Availability [14] Y. Aviv and A. Pazgal, “Optimal pricing of seasonal products
in the presence of forward-looking consumers,”
The raw/processed data required to reproduce the findings Manufacturing & Service Operations Management, vol. 10,
in this paper cannot be shared at this time as the data also no. 3, pp. 339–359, 2008.
form part of an ongoing study. [15] G. P. Cachon and R. Swinney, “The value of fast fashion: quick
response, enhanced design, and strategic consumer behavior,”
Management Science, vol. 57, no. 4, pp. 778–795, 2011.
Conflicts of Interest [16] P.-S. Chow, T.-M. Choi, and T. Cheng, “Impacts of minimum
order quantity on a quick response supply chain,” IEEE
The authors declare that there are no conflicts of interest Transactions on Systems, Man, and Cybernetics-Part A: Sys-
regarding the publication of this study. tems and Humans, vol. 42, no. 4, pp. 868–879, 2012.
10 Mathematical Problems in Engineering

[17] T.-M. Choi, “Local sourcing and fashion quick response

system: the impacts of carbon footprint tax,” Transportation
Research Part E: Logistics and Transportation Review, vol. 55,
pp. 43–54, 2013.
[18] H.-L. Chan, T.-M. Choi, C.-L. Hui, and S.-F. Ng, “Quick
response healthcare apparel supply chains: value of RFID and
coordination,” IEEE Transactions on Systems, Man, and Cy-
bernetics: Systems, vol. 45, no. 6, pp. 887–900, 2014.
[19] T.-M. Choi, “Inventory service target in quick response
fashion retail supply chains,” Service Science, vol. 8, no. 4,
pp. 406–419, 2016.
[20] Ö. Cosgun, U. Kula, and C. Kahraman, “Markdown opti-
mization for an apparel retailer under cross-price and initial
inventory effects,” Knowledge-Based Systems, vol. 120,
pp. 186–197, 2017.
[21] H.-M. Wee, “Economic production lot size model for dete-
riorating items with partial back-ordering,” Computers &
Industrial Engineering, vol. 24, no. 3, pp. 449–458, 1993.
[22] G. Li, X. He, J. Zhou, and H. Wu, “Pricing, replenishment and
preservation technology investment decisions for non-in-
stantaneous deteriorating items,” Omega, vol. 84, pp. 114–126,
2019.
[23] J. Zhang, Y. Wang, L. Lu, and W. Tang, “Optimal dynamic
pricing and replenishment cycle for non-instantaneous de-
terioration items with inventory-level-dependent demand,”
International Journal of Production Economics, vol. 170,
pp. 136–145, 2015.
[24] K.-S. Wu, L.-Y. Ouyang, and C.-T. Yang, “Coordinating re-
plenishment and pricing policies for non-instantaneous de-
teriorating items with price-sensitive demand,” International
Journal of Systems Science, vol. 40, no. 12, pp. 1273–1281, 2009.
[25] T. Avinadav, A. Herbon, and U. Spiegel, “Optimal inventory
policy for a perishable item with demand function sensitive to
price and time,” International Journal of Production Eco-
nomics, vol. 144, no. 2, pp. 497–506, 2013.
[26] A. Herbon and E. Khmelnitsky, “Optimal dynamic pricing
and ordering of a perishable product under additive effects of
price and time on demand,” European Journal of Operational
Research, vol. 260, no. 2, pp. 546–556, 2017.
[27] M. Fisher, K. Rajaram, and A. Raman, “Optimizing inventory
replenishment of retail fashion products,” Manufacturing &
Service Operations Management, vol. 3, no. 3, pp. 230–241,
2001.
[28] K. Kogan and U. Spiegel, “Optimal policies for inventory
usage, production and pricing of fashion goods over a selling
season,” Journal of the Operational Research Society, vol. 57,
no. 3, pp. 304–315, 2006.
[29] Y.-C. Tsao, “Two-phase pricing and inventory management
for deteriorating and fashion goods under trade credit,”
Mathematical Methods of Operations Research, vol. 72, no. 1,
pp. 107–127, 2010.