Available online at [Link].

com
Available online at [Link]

ScienceDirect
ScienceDirect
Procedia
Available Computer
online Science 00 (2023) 000–000
at [Link]
Procedia Computer Science 00 (2023) 000–000
[Link]/locate/procedia
[Link]/locate/procedia
ScienceDirect
Procedia Computer Science 232 (2024) 1233–1241

5th International Conference on Industry 4.0 and Smart Manufacturing

5th International Conference on Industry 4.0 and Smart Manufacturing
Perishable
Perishable Product
Product Inventory
Inventory Management
Management In In The
The Case
Case Of
Of Discount
Discount
Policies And Price-Sensitive Demand: Discrete Time Simulation
Policies And Price-Sensitive Demand: Discrete Time Simulation
And
And Sensitivity
Sensitivity Analysis
Analysis
Federico Solaria,a,*, Natalya Lysovaaa, Michele Bocelliaa, Andrea Volpiaa
Federico Solari *, Natalya Lysova , Michele Bocelli , Andrea Volpi
and Roberto Montanariaa
and Roberto Montanari
Department of Engineering and Architecture, University of Parma, Parco Area delle Scienze 181/A, 43124, Parma, Italy
Department of Engineering and Architecture, University of Parma, Parco Area delle Scienze 181/A, 43124, Parma, Italy

Abstract
Abstract
In B2C contexts, the demand is normally shifted towards products with higher residual shelf life (LEFO issuing policy). In such
In B2C contexts,
contexts, the risk isthethat
demand is normally
products closer to shifted towardsdate
the expiration products
could with
not behigher
pickedresidual shelf life (LEFO
up by consumers, issuing policy).
thus reaching the end In such
of their
contexts, the risk is that products closer to the expiration date could not be picked up by consumers, thus reaching
shelf life unsold, with the need to be disposed of. To reduce the extent of this phenomenon, retailers can promote the sale of expiring the end of their
shelf life unsold,
products with theappropriate
by introducing need to be disposed
discountof. To reduce
policies. the extent
In this paper,ofa this phenomenon,
multi-period retailers can
discrete-time promotemodel
simulation the saleis of expiring
developed
products
consideringby perishable
introducingproducts,
appropriate
havingdiscount
a fixedpolicies. In this paper,
and deterministic shelfa life,
multi-period discrete-time
issued according simulation
to a periodic reviewmodel is developed
policy. A mixed
considering
LEFO-FEFOperishable products,
issuing policy was having a fixed
considered. and deterministic
Specifically, shelf life,
in the absence of issued
discountaccording
policies,tothe
a periodic
demand review
is fulfilledpolicy. A mixed
according to
LEFO-FEFO issuing policy was considered. Specifically, in the absence of discount policies, the demand
LEFO logic. When a discount policy is applied, a portion of the demand, proportional to the discount, shifts to discounted expiringis fulfilled according to
LEFO logic. When a discount policy is applied, a portion of the demand, proportional to the discount,
products and is handled according to FEFO logic. The discount policy was defined as a percentage discount, applied to productsshifts to discounted expiring
products and is shelf
with a residual handledlifeaccording
of less thanto FEFO logic.
a defined The discount
number of days. policy was defined
The presented modelas was
a percentage
applied to discount,
a specific applied to products
case study and a
with a residual
sensitivity shelfwas
analysis lifeperformed
of less than to aassess
definedthenumber
impact of
of days. Thediscount
both the presented model
policy andwas applied
other contourto avariables,
specific case suchstudy and a
as demand
sensitivity analysis was performed to assess the impact of both the discount policy and other contour variables,
parameters (mean and standard deviation), lead time (LT) and product shelf life (SL), on average daily profit (𝑃𝑃). The results show � such as demand
parameters
that, for the(mean
context and standardthe
analysed, deviation),
introductionleadoftime (LT) and
a proper product
discount shelfleads
policy life (SL),
to anon average
increase indaily profit (𝑃𝑃�).0.31%
𝑃𝑃� of between The results show
and 2.45%.
that, for theitcontext analysed, �
Moreover, was obtained thatthe introduction
demand of a proper
variability discount
positively policy
affects leads toofandiscount
the impact increasepolicies.
in 𝑃𝑃 of between
Sensitivity0.31% and 2.45%.
analysis finally
Moreover,
showed thatit𝑃𝑃�was obtained that
is negatively demand
correlated variability
with positively correlated
LT and positively affects thewith
impact
SL. of discount policies. Sensitivity analysis finally
showed that 𝑃𝑃� is negatively correlated with LT and positively correlated with SL.
© 2024
© 2023TheTheAuthors.
Authors. Published
Published by ELSEVIER
by Elsevier B.V. B.V. This is an open access article under the CC BY-NC-ND license
© 2023 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license
([Link]
This is an open access article under the CC BY-NC-ND license ([Link]
([Link]
Peer-reviewunder
Peer-review underresponsibility
responsibility
ofof
thethe scientific
scientific committee
committee of5th
of the theInternational
5th International Conference
Conference on Industry
on Industry 4.0 and4.0 andManufacturing
Smart Smart
Peer-review
Manufacturingunder responsibility of the scientific committee of the 5th International Conference on Industry 4.0 and Smart
Manufacturing
Keywords: inventory management; periodic review; perishable products; price-sensitive demand; discount; discrete time simulation
Keywords: inventory management; periodic review; perishable products; price-sensitive demand; discount; discrete time simulation

* Corresponding author. Tel.: +39 0521905387; fax: +39 0521905705.

* Corresponding author.
E-mail address: Tel.: +39 0521905387; fax: +39 0521905705.
[Link]@[Link]
E-mail address: [Link]@[Link]
1877-0509 © 2023 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license
1877-0509 © 2023 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license
([Link]
([Link]
Peer-review under responsibility of the scientific committee of the 5th International Conference on Industry 4.0 and Smart Manufacturing
Peer-review under responsibility of the scientific committee of the 5th International Conference on Industry 4.0 and Smart Manufacturing

1877-0509 © 2024 The Authors. Published by Elsevier B.V.

This is an open access article under the CC BY-NC-ND license ([Link]
Peer-review under responsibility of the scientific committee of the 5th International Conference on Industry 4.0 and
Smart Manufacturing
10.1016/[Link].2024.01.121
1234 Federico Solari et al. / Procedia Computer Science 232 (2024) 1233–1241
2 Solari et al./ Procedia Computer Science 00 (2023) 000–000

1. Introduction

One of the 17 Sustainable Development Goals that the United Nations are targeting for 2030, specifically Goal
12.3, is to halve global per capita food waste at the retail and consumer level and reduce food losses during production
and supply chains, including post-harvest losses. In fact, it is estimated that about one-third of all food produced for
human consumption is wasted each year [1] and, with it, all the resources that were needed to produce it. According
to the Food and Agriculture Organization, it is necessary to make a distinction between food loss and food waste. The
former represents the waste that occurs in the first links of the production chain, while the latter refers to the last links
of the chain i.e. retail and consumption. In developing countries, most food is wasted during the intermediate stages
of production or due to storage problems; in rich countries, on the other hand, a large proportion of food is wasted in
the last stages of the distribution chain, namely by consumers and retailers. It is estimated that 13.8% of food is wasted
between collection and retail, the latter excluded, and 17% of food is wasted between retail, food services, and
consumption [2,3].
In retailing and distribution, the main causes of perishable product waste are overstocking, consumer behaviour,
and inappropriate quality control and product handling [4]. Both stockout and the disposal of perishable products
resulting from overstocking, as well as posing a problem in terms of sustainability, involve a cost for the retailer:
reducing waste in this area is, therefore, a goal to be pursued, both for environmental and profit reasons. It is therefore
necessary to identify an inventory management strategy that can optimise profit while minimizing waste. Several
studies can be found in the literature that have addressed this issue. In [5], the authors focused on the level of service
while neglecting the economic aspect of a periodic review policy. They assumed fixed and deterministic procurement
lead time and shelf life, negligible order issuing cost, variable demand, and reorder interval equal to one day. In this
context, the authors demonstrated that a dynamic reorder policy yields better results, in terms of service level, than
two simpler heuristic policies (base stock and constant order). In a subsequent study [6], the authors showed that in
cases where the order issuing cost cannot be neglected, a reorder interval greater than one day is economically
advantageous. In both above studies, demand is considered independent of the selling price and the age of the products.
In [7] authors conducted a stochastic dynamic optimization to identify the optimal joint pricing and production
schedule which maximise the total discounted profit of an inventory control policy of perishable products. Similarly,
in [8] the authors assumed the per unit selling price and the length of the replenishment cycle as decision variables to
maximise the profit of an inventory system for deteriorating items with price-sensitive demand. In [9] the authors
considered a price and time-dependent ramp-type demand function and developed a mathematical model to maximise
the profit based on the number of price changes and the length of the replenishment cycle. In all these latter studies, it
is assumed that customers do not differentiate between products of different ages. In [10] the authors evaluate the
performance of two replenishment policies by considering differential demand according to product age and assessing
the impact of substitution when the requested product is not available (i.e., satisfying demand with a product having a
different age than the requested one). To the best of our knowledge, there are still no studies in the literature that refer
to the specific context of the retail store, where demand is highly dependent both on the selling price but also on the
presence of products having different ages. Indeed, in the retail environment, it is well known that products with a
close expiration date are perceived by consumers as having lower quality, demand therefore tends to prefer products
having a longer remaining shelf life [11]. Products closer to expiration, therefore, remain on the shelves and, as they
reach the end of their shelf life, must be disposed of. The application of discounts has been found in the literature to
have an important effect on consumer behaviour [12]. The application of discounting policies helps to bring the
consumer's attention back to products closer to expiration by reducing waste and contributing to profit [13].
In this paper, a discrete-time simulation model was built to simulate the operating cycle of a warehouse for
perishable products, which involves a periodic review inventory policy, i.e., orders are issued at regular time intervals
(T) and a quantity of products is ordered such that a maximum level (S) is restored. A single-product scenario was
considered, and demand, having normal distribution, was assumed to be fulfilled according to a mixed Last-Expired-
First-Out (LEFO)—First-Expired-First-Out (FEFO) logic. Specifically, in the absence of discount policies, all demand
is met according to LEFO logic. When a discount policy is applied, a portion of the daily demand, proportional to the
discount applied, shifts to discounted expiring products and is handled according to FEFO logic. The discount policy
was defined based on two parameters: a discount percentage (%D) and a residual shelf life (RSL), from which, the
product is sold at a discounted price, until the end of its useful life.
 Federico Solari et al. / Procedia Computer Science 232 (2024) 1233–1241 1235
Solari et al./ Procedia Computer Science 00 (2023) 000–000 3

The model was then applied to some case studies characterised by different values of the mean (µ) and standard
deviation (σ) of demand, procurement lead time (LT), and product shelf life (SL). The optimal configuration, defined
as a combination of the operating parameters (T, S, %D and RSL) that maximises average daily profit, was identified
through a grid-search optimization method. The impact that discounting policies have on average daily profit was
evaluated, and a sensitivity analysis was also conducted to assess the impact that procurement lead time and product
shelf life have on system performance.
Many works can be found in the literature on the identification of the optimal discount policy to maximise profit,
and comprehensive reviews on this topic can be found as well [12,14]. Authors in [13] focused on the combined effect
that discount policies and dynamic shelf life have on performance in terms of waste reduction, profit, shortages, and
product quality. None of the studies mentioned assessed how a change in demand parameters, supply lead time or
product shelf life impacted the system performance. The proposed work therefore aims to fill this gap by assessing the
impact that variable parameters have on optimal operational parameters and management policy performance.
The outline of the article is as follows: Section 2 describes the methodology and the assumptions adopted and
introduces the case study analysed with the relative optimization procedure. In Section 3, the results obtained are
presented and discussed. In Section 4 some managerial insights, derived from the application of your simulation model,
are reported, and in Section 5 conclusions, limitations of the work and insights for future activities are presented.

2. Materials and methods

2.1. Overview and assumptions

The context studied is a retail store that manages the stock of perishable products, having a deterministic shelf life,
by means of a periodic reorder policy, considering a positive, deterministic procurement lead time.
A multi-period, single-product, discrete-time simulation model was developed that reproduces the warehouse
working cycle. A single day was chosen as the reference period and 5000 periods were considered, during which the
flows of products in the warehouse were observed.
The main assumptions are as follows:
- Products are characterised by normally distributed demand, having known mean and standard deviation;
- Demand is satisfied according to the LEFO logic, which is transformed into a mixed LEFO-FEFO logic
when a certain discount percentage is applied on products having a remaining shelf life equal to or lower
than a certain residual shelf life (RSL).In particular, when a discount is applied, a percentage of demand,
equal to the discount applied, shifted shifts to discounted expiring products and is handled according to
FEFO logic.
- Products are characterised by a positive and deterministic shelf life;
- Products retain their value until the end of their useful life;
- A maximum warehouse capacity of 24000 units is assumed;
- Orders are issued at regular periods and, at each reorder, a quantity of products is ordered to reestablish a
given level of stock in the warehouse;
- The target stock level is assumed to be a multiple of 500;
- The reorder period is assumed to be shorter than the shelf life;
- All products belonging to the same order have the same shelf life value, which begins to decrease the
moment the product is shipped;
- The procurement lead time is considered positive and deterministic;
- Within the warehouse, products are grouped according to the order in which they arrive at the warehouse,
and then according to their residual shelf life;
- Shortages are admitted and fully back-ordered;
Finally, the following economic aspects are accounted for within the model: inventory holding cost, stock-out cost,
order issuing cost, disposal cost, purchasing cost, sales revenue, and profit. The order issuing cost is assumed to be
independent of the ordered quantity.
1236 Federico Solari et al. / Procedia Computer Science 232 (2024) 1233–1241
4 Solari et al./ Procedia Computer Science 00 (2023) 000–000

Nomenclature

i i-th day (i=1….n)

di Daily demand
µi Average daily demand
σi Standard deviation of daily demand
LT Procurement Lead Time
T Reorder period
S Target stock level
SL Shelf life
OHi Stock level at day i
Oi Quantity purchased at day i
OOSi Out-of-stock at day i
cinv Unitary inventory holding cost
cso Unitary stock-out cost
cdisp Unitary disposal cost
cp Unitary purchase cost
Coi Order issuing cost
p selling price
R Sales revenue
P Profit
𝑃𝑃� Average daily profit
RSL Residual shelf life value from which a discount is applied
%D Discount percentage
Qdisp,i Disposed quantity at day i
Qexp,i Expired quantity at day i
Qs,i quantity sold at full price at day i
qs,i quantity sold at a discounted price at day i

Fig. 1 shows the logic of the developed model, which are described as follows: for each period i the following
activities are performed (Fig. 1):
- products that have reached the end of their shelf life (𝑄𝑄���,� ) are disposed of and the stock-on-hand (𝑂𝑂𝑂𝑂� )
is updated accordingly;

𝑂𝑂𝑂𝑂� = 𝑂𝑂𝑂𝑂��� − 𝑄𝑄���,� (1)

- products that may have been ordered LT days in advance are received in stock and the stock-on-hand is
updated accordingly;

𝑂𝑂𝑂𝑂� = 𝑂𝑂𝑂𝑂� + 𝑂𝑂���� (2)

- Any unfulfilled demand from previous days, if sufficient stock is available, is fulfilled (backorder);

𝑂𝑂𝑂𝑂� = 𝑂𝑂𝑂𝑂� − 𝑂𝑂𝑂𝑂𝑂𝑂��� (3)

- According to the periodic review policy, if an order is scheduled for period i, a quantity such as to restore
a stock level equal to S is issued. The related cost (Coi,i) which is fixed and independent of Oi is computed;

𝑂𝑂� = 𝑆𝑆 − 𝑂𝑂𝑂𝑂� (4)

 Federico Solari et al. / Procedia Computer Science 232 (2024) 1233–1241 1237
Solari et al./ Procedia Computer Science 00 (2023) 000–000 5

Fig. 1. Flowchart of the proposed approach.

- For products in stock having a remaining shelf life equal or lower than RSL, a discount equal to %D is
applied;
- if discounted products are present in stock, a percentage of daily demand, equal to %D, shifts towards
these products. The remaining portion of demand, on the other hand, remains oriented to products with
the longest remaining shelf life;
- In both cases, if the stock-on-hand doesn’t meet the demand an out-of-stock situation arises.

𝑂𝑂𝑂𝑂𝑂𝑂� = 𝑂𝑂𝑂𝑂𝑂𝑂��� + 𝑑𝑑� − 𝑂𝑂𝑂𝑂� (5)

The main cost items were then computed according to the following equations:

𝐶𝐶���,� = 𝑐𝑐��� ∙ 𝑂𝑂𝑂𝑂� (6)

𝐶𝐶��,� = 𝑐𝑐�� ∙ 𝑂𝑂𝑂𝑂𝑂𝑂� (7)

𝐶𝐶����,� = 𝑐𝑐���� ∙ 𝑄𝑄���,� (8)

𝐶𝐶�,� = 𝑐𝑐� ∙ 𝑂𝑂� (9)

𝐶𝐶���,� = 𝐶𝐶���,� + 𝐶𝐶��,� + 𝐶𝐶����,� + 𝐶𝐶��,� + 𝐶𝐶�,� (10)

Contextually, sales-related revenues and resulting profit are also calculated.

𝑅𝑅� = 𝑄𝑄�,� ∙ 𝑝𝑝 𝑝 𝑝𝑝�,� ∙ 𝑝𝑝 𝑝 𝑝𝑝𝑝 (11)

𝑃𝑃� = 𝑅𝑅� − 𝐶𝐶���,� (12)

 1238 Federico Solari et al. / Procedia Computer Science 232 (2024) 1233–1241
6 Solari et al./ Procedia Computer Science 00 (2023) 000–000

By repeating this process for a sufficiently large number of periods (5000 in this study), the expected outcomes
that the inventory management policy achieves can be calculated. In this study, the average daily profit was considered
as the main result:

∑ �
�
𝑃𝑃� = ��� � (13)
�

2.2. Numerical case study and sensitivity analysis

To quantify the impact that some of the variables involved have on system performance, a sensitivity analysis was
performed. As variables were selected those over which the inventory manager has limited influence, but which
influence his choices: demand parameters, procurement lead time and product shelf life. For each variable three levels
were considered, then both the minimum, the maximum and the mean value to be investigated were defined. The
unitary costs as well as the selling price were considered to be fixed (cinv= 0.024 €/unit, cso= 0.47 €/unit, cdisp= 0.163
€/unit, cp= 5 €/unit, Coi= 550 €/order, p= 8 €/unit)A total of nine case studies, as summarised in Table 1, were
considered.

Table 1. Case studies analysed.

Parameter CS 1 CS 2 CS 3 CS 4 CS 5 CS 6 CS 7 CS 8 CS 9 Unit
µi 2500 2000 3000 2500 2500 2500 2500 2500 2500 Units
σi 500 500 500 400 600 500 500 500 500 units
LT 2 2 2 2 2 1 3 2 2 Days
SL 10 10 10 10 10 10 10 8 12 Days

2.3. Optimization procedure

The optimization procedure was performed in two successive steps. In the first step, the reorder policy was
optimised by identifying the combination of operational parameters T and S that would minimise the total daily cost.
Next, the optimal discount policy was identified for each configuration, defined as the combination of %D and RSL
that maximised the average daily profit. A grid search-based optimization method was used. In Table 2 the variation
ranges of each operational parameter were defined. Each parameter was then varied within that range and all possible
combinations were evaluated to identify the optimal one.

Table 2. Variation ranges of the operating parameters.

Parameter Min value Max value Increasing step
T 2 SL 1
S 8000 24000 500
CSL 1 3 1
%D 0 0.5 0.1

3. Results and discussion

Fig. 2 shows the results related to the policy optimization without considering any discount policy, i.e., as if it were
a standard reorder period policy, in the case of SL=10 and LT=2. The maximum average daily profit, equal to 6674.80
€, is obtained at T=4 and S=13500.
 Solari etSolari
Federico al./ Procedia ComputerComputer
et al. / Procedia Science 00 (2023)232
Science 000–000
(2024) 1233–1241 12397

Fig. 2. Average daily profit in function of T and S, without considering any discount policy

Fig. 3. Impact of the discount policy on the average daily profit

Results highlight that, for the case study considered, applying a discount at the last day of useful life has a less
significant impact than applying it with 2 or 3 days of residual shelf life. It can also be seen that for high discount
percentages, the average daily profit trends downward: in this area, the lost profits caused by selling at a discounted
price outweigh the reduction in the disposal and stock-out costs. Table 3 displays the results related to the optimal
discount policies of the nine investigated case studies. Finally, the impacts that the parameters considered have on
system performance with and without discount policy were evaluated. The results are depicted in Fig. 4.
As might be expected, lead time is negatively correlated with profit, i.e., an increase in LT causes a decrease in 𝑃𝑃� ,
and vice versa. On the other hand, the correlation between SL and 𝑃𝑃� is positive and the impact of the discount policy
is higher the shorter is SL. Moreover, as average demand increases, the impact of the discount policy slightly increases:
the percentage increase in profit therefore is greater when, in the presence of discounting, average demand is higher.
Regarding demand variability, it emerges how the impact of discount policy increases as the standard deviation of
demand increases. Finally, the impact of discounting decreases as the procurement lead time increases, because for
longer lead times, the on-shelf exposure time of the product decreases.
1240 Federico Solari et al. / Procedia Computer Science 232 (2024) 1233–1241
8 Solari et al./ Procedia Computer Science 00 (2023) 000–000

Table 3: Optimal reordering policies and optimal discount policies for the nine case studies analysed

Case T S RSL Average daily profit Average daily profit Impact of

%D
study [days] [units] [days] without discount [€] with discount [€] discount

CS 1 4 13500 20% 3 6674,80 6811,83 2,05%

CS 2 4 11000 20% 3 5190,05 5293,15 1,99%
CS 3 4 16000 20% 3 8202,91 8347,86 1,77%
CS 4 4 13500 20% 3 6825,90 6951,29 1,84%
CS 5 4 14000 20% 3 6651,25 6800,13 2,24%
CS 6 2 6000 10% 3 6793,82 6946,10 2,24%
CS 7 6 21000 10% 1 6504,74 6524,94 0,31%
CS 8 3 11000 10% 3 6516,92 6676,31 2,45%
CS 9 5 16500 20% 3 6736,17 6866,16 1,93%

a b

c d

Fig. 4. Impact of mean of demand (a), demand standard deviation (b), Shelf life (c) and procurement lead time (d) on the average daily profit with
(red) and without (blue) discount policy

4. Managerial insights

The following managerial insights can be deduced from the study conducted:
 The impact of discount policy, in terms of increasing average daily profit, varies with system parameters.
 The increase in daily profit is between 0.31% and 2.45% (without considering any additional demand
generated by the introduction of the discount).
 The impact of the discount is greatest when the variability of demand is high.
 The impact of the discount is highest when the supply lead time is short.

5. Conclusions

In B2C contexts, the shelf-picking process, in the absence of discount policies, follows a LEFO logic, as consumers
believe that products which are farther from expiration have higher quality. This means that the more time passes, the
greater the probability that products with shorter residual shelf lives remain on the shelf. The risk that such products
 Federico Solari et al. / Procedia Computer Science 232 (2024) 1233–1241 1241
Solari et al./ Procedia Computer Science 00 (2023) 000–000 9

will have to be disposed of therefore increases. To reduce this risk, the retailer can introduce discounts to encourage
consumers to purchase expiring products.
In this study, a discrete-time simulation model to evaluate the impact of discount policies in a B2C context is
presented. The model was applied to nine case studies and a sensitivity analysis was performed to assess the impact
that both the discount policy as well as other contour variables, such as demand features, lead time and shelf life of
products, have on average daily profit. It was found that, for the case studies investigated, the introduction of an
appropriate discount policy can lead to an increase in average daily profit between 0.31% and 2.45%. It was also found
that the optimal discount percentage depends on the remaining shelf life value from which the discount is applied. In
most cases considered, the optimal discount policy was found to be the one consisting of a 20% discount applied to
products having a remaining shelf life of three days.
In future research, it might be interesting to further investigate the sensitivity analysis by designing a simulative
campaign using the Design of Experiments methodology and performing a statistical analysis of the results to quantify
the relationships that exist among the variables in the system. Also, the discount percentage could be correlated to the
product's residual shelf life. The presented model has limitations that can be addressed in future research activities.
For example, the demand is assumed to have a fixed statistical distribution. Other demand characteristics can be
studied such as seasonality and multi-modality, as well as demand trends can be derived through data mining
techniques based on time series. The assumption that shortages are fully backordered can also be considered a
limitation, since in a B2C context, product shortages often result in lost sales. A context in which shortages are partly
lost and partly back-ordered may be considered in future research activities. Moreover the assumption, already found
in the literature, that the percentage discount introduced displaces a percentage of demand, equal to the percentage of
discount applied, on expiring products, also represents a limitation. This assumption will need to be investigated with
dedicated studies to further explore the relationship that exists between discount policies and consumer behaviour,
also depending on the context.

References

[1] FAO, G. (2011). Global food losses and food waste-Extent, causes and prevention. SAVE FOOD: An initiative on food loss and waste reduction,
9, 2011
[2] FAO, I. (2019). The state of food and agriculture 2019. Moving forward on food loss and waste reduction. FAO, Rome, 2-13
[3] UNEP, 2021. Food Waste Index Report 2021 (Nairobi)
[4] Wang, X., & Li, D. (2012). A dynamic product quality evaluation based pricing model for perishable food supply chains. Omega, 40(6), 906–
917. [Link]
[5] Minner, S., Transchel, S., (2010). Periodic review inventory-control for perishable products under service-level constraints. OR Spectrum 32,
979–996. [Link]
[6] Kouki, C., Jemai, Z., Sahin, E., Dallery, Y., (2014). Analysis of a periodic review inventory control system with perishables having random
lifetime. Int J Prod Res. [Link]
[7] Li, S., Zhang, J., & Tang, W. (2014). Joint dynamic pricing and inventory control policy for a stochastic inventory system with perishable
products. Int J Prod Res., 53(10), 2937–2950. [Link]
[8] Khan, Md. A.-A., Ahmed, S., Babu, Md. S., & Sultana, N. (2020). Optimal lot-size decision for deteriorating items with price-sensitive demand,
linearly time-dependent holding cost under all-units discount environment. International Journal of Systems Science: Operations &
Logistics, 9(1), 61–74. [Link]
[9] Panda, S., Saha, S., & Basu, M. (2013). Optimal pricing and lot-sizing for perishable inventory with price and time dependent ramp-type demand.
International Journal of Systems Science, 44(1), 127–138. [Link]
[10] Deniz, B., Karaesmen, I., & Scheller-Wolf, A. (2020). A comparison of inventory policies for perishable goods. Operations Research Letters,
48(6), 805–810. [Link] Tsiros, M., & Heilman, C. M. (2005). The Effect of Expiration Dates and
Perceived Risk on Purchasing Behavior in Grocery Store Perishable Categories. Journal of Marketing, 69(2), 114–129.
[Link]
[12] Elmaghraby, W., & Keskinocak, P. (2003). Dynamic Pricing in the Presence of Inventory Considerations: Research Overview, Current
Practices, and Future Directions. Management Science, 49(10), 1287–1309. [Link]
[13] Buisman, M. E., Haijema, R., & Bloemhof-Ruwaard, J. M. (2019). Discounting and dynamic shelf life to reduce fresh food waste at retailers.
International Journal of Production Economics, 209, 274–284. [Link]
[14] Bakker, M., Riezebos, J., & Teunter, R. H. (2012). Review of inventory systems with deterioration since 2001. European Journal of Operational
Research, 221(2), 275–284. [Link]