systems

Article
Strategies for Humanitarian Logistics and Supply Chain in
Organizational Contexts: Pre- and Post-Disaster
Management Perspectives
Amir Aghsami 1 , Simintaj Sharififar 2,3, *, Nader Markazi Moghaddam 4,5 , Ebrahim Hazrati 6 , Fariborz Jolai 7
and Reza Yazdani 8

1 Trauma and Surgery Research Center, Aja University of Medical Sciences, Tehran 14117-18541, Iran;
a.aghsami@ut.ac.ir
2 Toxicology Research Center, Aja University of Medical Sciences, Tehran 14117-18541, Iran
3 School of Nursing, Aja University of Medical Sciences, Tehran 14117-18541, Iran
4 Cancer Epidemiology Research Center, Aja University of Medical Sciences, Tehran 14117-18541, Iran;
nmmoghaddam@sbmu.ac.ir
5 Critical Care Quality Improvement Research Center, Shahid Modarres Hospital, Shahid Beheshti University
of Medical Sciences, Tehran 14117-18541, Iran
6 Infections Disease Research Center, Aja University of Medical Sciences, Tehran 14117-18541, Iran;
dr.hazrati.e@ajaums.ac.ir
7 School of Industrial and Systems Engineering, College of Engineering, University of Tehran,
Tehran 14117-18541, Iran; fjolai@ut.ac.ir
8 Department of Accounting, Technical and Vocational University (TVU), Tehran 14117-18541, Iran;
r-yazdani@tvu.ac.ir
* Correspondence: s.sharififar@ajaums.ac.ir

Abstract: Every organization typically comprises various internal components, including regional
branches, operations centers/field offices, major transportation hubs, and operational units, among
others, housing a population susceptible to disaster impacts. Moreover, organizations often possess
resources such as staff, various vehicles, and medical facilities, which can mitigate human casualties
Citation: Aghsami, A.; Sharififar, S.; and address needs across affected areas. However, despite the importance of managing disasters
Markazi Moghaddam, N.; Hazrati, E.; within organizational networks, there remains a research gap in the development of mathematical
Jolai, F.; Yazdani, R. Strategies for models for such scenarios, specifically incorporating operations centers/field offices and external
Humanitarian Logistics and Supply stakeholders as relief centers. Addressing this gap, this study examines an optimization model for
Chain in Organizational Contexts: Pre-
both before and after disaster planning in a humanitarian supply chain and logistical framework
and Post-Disaster Management
within an organization. The affected areas are defined as regional branches, operational units, major
Perspectives. Systems 2024, 12, 215.
transportation hubs, operations centers/field offices, external stakeholders, and medical facilities.
https://doi.org/10.3390/
A mixed-integer nonlinear model is formulated to minimize overall costs, considering factors such
systems12060215
as penalty costs for untreated injuries and demand, delays in rescue and relief item distribution
Academic Editors: Shuqi Xue, operations, and waiting costs for the injured in emergency medical vehicles and air ambulances.
Yun Wang and Xiaomeng Shi
The model is implemented using GAMS software 47.1.0 for various test problems across different
Received: 12 May 2024 scales, with the Grasshopper Optimization Algorithm proposed for larger-scale scenarios. Numerical
Revised: 9 June 2024 examples are provided to show the effectiveness and feasibility of the proposed model and to validate
Accepted: 13 June 2024 the metaheuristic approach. Sensitivity analysis is conducted to assess the model’s performance
Published: 18 June 2024 under different conditions, and key managerial insights and implications are discussed.

Keywords: humanitarian logistics; supply chain; organizations; disaster management; staffs; relief
items; distribution operations
Copyright: © 2024 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
1. Introduction
Attribution (CC BY) license (https:// Throughout history, humanity has faced a multitude of natural disasters, including
creativecommons.org/licenses/by/ floods, earthquakes, hurricanes, and more, resulting in serious and often irreversible con-
4.0/). sequences despite technological advancements in recent decades [1–3]. Some of these

Systems 2024, 12, 215. https://doi.org/10.3390/systems12060215 https://www.mdpi.com/journal/systems

Systems 2024, 12, 215 2 of 29

disasters have been so catastrophic that they have significantly impacted the fundamental
structures of countries, including their economies, societies, physical infrastructure, and,
most importantly, the lives of those affected in densely populated areas [4–6]. Natural dis-
asters have been recurring events, causing numerous injuries across various regions, with
the frequency of occurrences such as tsunamis and earthquakes increasing over time [7,8].
Consequently, casualties have risen significantly due to socio-economic and demographic
factors such as urban development and population growth [9]. Over the last century, an
immense loss of human life has been attributed to natural disasters, with over 2.4 million
fatalities and more than 206 million individuals injured, displaced, or left unemployed
between 1900 and 2020 [10]. Moreover, sudden disasters have impacted billions of peo-
ple worldwide in the past decade [11]. For instance, devastating cyclones in Bangladesh
(Cyclone Bhola, 1970), the Philippines (Typhoon Haiyan, 2013), India (Cyclone Odisha,
1999), Myanmar (Cyclone Nargis, 2008), and the United States (Hurricane Katrina, 2005)
collectively led to the loss of an estimated 400,000 lives. A notable example occurred
when a powerful earthquake measuring 6.6 Mw struck the Iburi Subprefecture in south-
ern Hokkaido, Japan, causing numerous injuries and widespread damage, including the
disruption of electrical services throughout Hokkaido. In response to such devastating
events, it becomes imperative to provide an effective disaster response, with particular
emphasis on the humanitarian logistics and supply chain aimed at minimizing casualties
and damage in affected areas [12,13].
The humanitarian supply chain (HSC) encompasses various activities designed to
plan, execute, and manage the efficient flow and storage of inventory to minimize damage
and casualties during natural disasters [14,15]. Moreover, humanitarian logistics performs
a crucial role in reducing the consequences of disasters, and encompasses pre-disaster and
post-disaster actions. Pre-disaster operations, like establishing permanent warehouses and
prepositioning relief items in these facilities, aim to decrease the impact of disasters and
enhance the response time to demands in affected areas. Post-disaster operations involve
the distribution of relief items (RIs), the evacuation of injured individuals, and the transfer
of critically injured individuals to medical centers [16].
In essence, numerous large organizations encompass diverse components spanning
multiple geographical areas. By harnessing the inherent resources and capabilities of these
entities, they possess the capacity to mitigate crises to a significant degree, thereby minimiz-
ing both human and financial losses, especially in scenarios where external assistance to
affected areas within the organization is absent or delayed. Despite the extensive literature
on disaster management and humanitarian logistics and supply chain, scant attention has
been devoted to examining intra-organizational disaster management practices and the
utilization of existing organizational facilities in disaster preparedness and response efforts.
Moreover, a notable lack of optimization models exist to analyze disaster management
within organizations, considering all major internal components. Hence, there is a crucial
requirement to establish a reliable and suitable framework for disaster management within
organizations, encompassing various components such as regional branches, operational
units, operations centers/field offices, etc. Given that each organization comprises numer-
ous parts and centers located in different places, some of these centers may be susceptible
to disasters and require immediate response, while the organization itself possesses crisis
management capabilities, including helicopters, medical centers, and staff. To address this
main research gap, this article develops an optimization mathematical model for planning
before and after disasters in humanitarian supply chains and logistics within organizations.
This paper contributes to the existing literature in several significant ways. Firstly,
it presents a novel humanitarian supply chain and logistics model tailored for disaster
management within the internal network of organizations. To the best of the authors’
knowledge, no prior studies have proposed a mathematical model specifically addressing
the internal operations of organizations within the context of humanitarian logistics and
supply chain management. Secondly, this paper emphasizes the importance of optimizing
the utilization of an organization’s internal capabilities for efficient disaster management.
Systems 2024, 12, 215 3 of 29

Consequently, the model incorporates an assessment of the organization’s internal resources

and facilities, including operations centers or field offices, external stakeholders, and
helicopters. It is evident that staff members stationed at operations centers/field offices
and external stakeholders can play a crucial role in rescue operations and distributing
RIs. Hence, the developed model contemplates the availability of staff within the network
to facilitate rescue operations and distribute relief items effectively. Thirdly, this paper
addresses critical challenges such as the waiting time for injured persons in emergency
medical vehicles (EMVs) or Air Ambulances (AAs) to reach medical centers, as well
as variations in the performance levels of rescue groups across various medical centers.
These considerations are integrated into the proposed mathematical model to enhance its
comprehensiveness and practical relevance.
This research is looking forward to answering the below questions:
• What is the best policy for managing disasters in the interior body of an organization?
• In which candidate sites, such as regional branches, operations centers/field offices,
external stakeholders, major transportation hubs, and operational units, should the
permanent relief centers (PRCs) be opened, and how many different RIs should be
propositioned in the established PRCs?
• How many staff should be transported from operations centers/field offices and
external stakeholders to affected areas for helping rescue activities and the distribution
of RIs separately?
• How many ground vehicles, helicopters, AAs, and EMVs should the organization
provide to effectively manage a disaster?
• How many untreated, injured, and unsatisfied demands would the organization have?
In order to answer the above questions, this paper designs an organization-based
humanitarian logistics and supply chain network within the organization network to
plan pre-and post-disaster operations. The affected areas within the organization, PRCs,
temporary relief centers (TRCs), operations centers/field offices, external stakeholders, and
medical centers are the main centers considered in the model. In addition, the affected
areas can be regional branches, major transportation hubs, operational units, operations
centers/field offices, external stakeholders, and medical centers in the proposed model.
The remainder of the paper is structured in the following way. A literature review is
discussed in Section 2. The problem description and mathematical model are presented
in Section 3. Section 4 discusses the solution methodology. Extensive numerical examples
are investigated and analyzed in Section 5. We present the sensitivity analysis in Section 6.
Section 7 expresses the discussions and some managerial implications. Eventually, the
conclusions are given in Section 8.

2. Literature Review
In this section, we provide an overview of the research background in the realm of
supply chain and humanitarian logistics in disaster management. It is noteworthy that
no study has delved into this area specifically concerning organizational contexts, a gap
we aim to address in the subsequent discussion. Given the surge in natural disasters
over the past two decades, numerous studies have explored this subject under various
circumstances. The recent literature has also undertaken reviews of papers within this
domain. For example, an overview of the HSC was discussed by Kovács and Spens [17],
and a summary of inventory management in HSC was conducted by Balcik et al. [18].
A significant part of the research in this area has developed the mathematical optimiza-
tion model under various conditions. Yi and Özdamar [19] presented a routing-location
model in the post-disaster phase that combines logistics support and the evacuation process.
Berkoune et al. [20] developed a mathematical model for the delivery of relief goods to
the affected areas, where the transportation time is minimized. Rawls and Turnquist [21]
investigated a model for pre-disaster emergency response that considers the location of
relief centers and various types of essential relief goods. A stochastic programming mathe-
matical model in an HSC aimed at minimizing establishment, inventory, transportation,
Systems 2024, 12, 215 4 of 29

and shortage costs was proposed by Döyen et al. [22]. The transferring of patients to
the medical center, along with the distribution of relief items to the affected areas, was
discussed in the article by [23]. One of their goals was to minimize the waiting time of
patients from the time of the accident to the moment they are taken to the medical center.
Camacho-Vallejo et al. [24] developed a humanitarian logistics model for international aid
distribution, with the aim of reducing transportation costs and increasing responsiveness.
Rezaei-Malek et al. [25] studied an integrated distribution and location-allocation model
in the pre-disaster stage to compute the optimal order quantity for the supply of RIs in
warehouses for perishable products. A humanitarian logistics network considering regional
distribution centers and warehouses in the crisis was developed by Tofighi et al. [26], who
examined the model in scenario-oriented terms. Another research that studied the HSC
with a scenario-based approach and uncertain demand was presented by Hu et al. [27].
Noham and Tzur [28] considered a humanitarian relief supply chain that distributed one
type of relief item to affected areas. Vahdani et al. [29] investigated a multi-objective, multi-
product, and multi-period two-stage mathematical model for the distribution of RIs. Liu
et al. [30] extended a resilience model for planning how to distribute relief items with the
aim of minimizing unmet demands after a disaster occurs. This study aims to determine
the location of warehouses, inventory management, and how to distribute goods in an
optimal condition.
Momeni et al. [31] proposed an HSC regarding the reliability of routes, repair groups,
and evaluation of routes. In the paper, demand points were prioritized based on population,
ensuring that relief goods are sent to demand points with higher priority. A model for
distributing and redistributing relief items in the post-disaster phase using a three-stage
systematic method was investigated by Sakiani et al. [32]. Also, to minimize the costs
and maximize the coverage of the damaged areas, a two-stage model was evaluated by
Aghajani et al. [33]. Heydari et al. [34] introduced a framework for debris clearance aimed at
reducing the time taken to reach critical sites such as medical centers, while optimizing the
extent of the area serviced. Abazari et al. [35] proposed a nonlinear multi-objective model
that considered relief centers, demand points, methods of distributing goods, and two types
of goods: perishable and non-perishable. Abazari et al. [36] analyzed a mathematical model
for designing an HSC in order to locate relief centers, distribute relief items, and provide
medical treatment in affected areas. A bi-objective optimization model was developed
for the first time by Hajipour et al. [37] to analyze an HSC combined with the traceability
concept. A multi-objective problem with multiple products and multiple periods in an HSC
considering trucks queue in the borders of the affected zones was studied by Masoumi
et al. [38], in which a routing-inventory-queue model was used to respond to demand
points with the aim of minimizing response time and reducing traffic and congestion.
Bakhshi et al. [39] modeled a relief supply chain problem in which non-governmental
organizations assist governmental organizations in relief activities. This model was devel-
oped for location-allocation, inventory management, and distribution planning of relief
goods under uncertainty for demand, budget, storage, and transportation costs. Ehsani
et al. [40] designed an HSC problem for a contagious disease, such as COVID-19, utilizing
the Internet of Things to identify infected individuals. Sheikholeslami and Zarrinpoor [41]
focused on a humanitarian logistics network design for response and preparedness stages
in times of crisis. Lohrasbpoor et al. [42] addressed a four-level humanitarian network,
where the degree of compatibility and destruction of blood units in the state of an earth-
quake was assessed. Khalili-Fard et al. [43] presented a comprehensive model for a pre-
and post-disaster humanitarian supply chain network, considering pre-positioning strat-
egy, NGO and GO collaboration in post-disaster distribution planning. Altay et al. [44]
presented a systematic literature review to specify gaps in the literature on innovation in
HSCs, and developed a suitable framework for future research. The goal of [45] is to asses
pivotal challenges for implementing the relief measures of humanitarian logistics in the
pre- and post-disaster phase. Beiki Ashkezari et al. [46] developed a bi-objective MINLP
model to consider pre-positioning and distributing relief items, which is formulated as
Systems 2024, 12, 215 5 of 29

an integrated location-allocation-routing problem under uncertainty. Mousavi et al. [47]

studied a combined mathematical and simulation model to design a hierarchical network
of temporary medical facilities during a disaster.
According to the review of related papers and the authors’ best knowledge, no study
has yet proposed a mathematical model specifically addressing disaster management in
a humanitarian logistics and supply chain framework with a focus on organizations and
considering all main internal components. Developing a comprehensive model is crucial for
organizations to successfully manage disasters within their own internal networks, leverag-
ing their existing facilities and staff. Therefore, this study introduces an organization-centric
approach to humanitarian logistics and supply chain planning, tailored for pre- and post-
disaster operations. In this paper, we examine the internal components of an organization,
including regional branches, operational centers/field offices, major transportation hubs,
external stakeholders, medical centers, and operational units. The definition of the organi-
zation in this study is quite comprehensive and covers various essential components. The
following is a breakdown of each part:
• Regional branches: These represent decentralized units or branches of the organization
located in different regions. They can help ensure that the organization has a presence
and can effectively serve its stakeholders across various geographic areas.
• Operational units: These are the frontline teams responsible for executing operational
tasks within the organization. They encompass different departments or teams in-
volved in core operational activities, such as production, service delivery, logistics,
and customer support.
• Major transportation hubs: These are critical facilities or locations where transportation
activities are centralized or concentrated. They play a crucial role in facilitating the
movement of goods, services, and people within the organization’s network.
• Operations centers/field Offices: These represent centralized or decentralized facilities
that serve as command and control centers for monitoring, coordinating, and man-
aging operational activities. They can include operations centers, command centers,
control rooms, and field offices, depending on their specific functions and locations.
• External stakeholders: These are the various entities outside the organization that
have a stake or interest in its activities, decisions, or outcomes. External stakeholders
can include suppliers, customers, partners, regulatory agencies, government bodies,
communities, and other organizations or individuals with whom the organization
interacts or collaborates.
• Medical centers: These represent facilities or units within the organization dedicated
to providing healthcare services. They can include medical centers, clinics, medical
centers, and other healthcare facilities that offer medical treatment, diagnostic services,
and patient care.
Overall, this definition covers a wide range of internal parts and external relationships
within the organization, which provides a comprehensive understanding of its structure,
operations, and interactions. Government agencies serve as prime examples of organiza-
tions, comprising various essential components such as regional or district offices, airports
managed by government entities, government-owned corporations, department of health
services, border security, public medical centers, or healthcare facilities operated by gov-
ernment entities, which are correspond to the abovementioned centers for an organization.
An additional illustration involves the utilization of a multinational logistics corporation’s
diverse infrastructure and resources, including regional branches, transportation hubs,
operational units, operations centers, external stakeholders, and medical centers. These
corporations often maintain regional branches across various affected regions and possess
transportation centers. Additionally, they have operational units, such as warehouses and
manufacturing facilities. Moreover, they have operations centers with numerous personnel
to coordinate logistical activities, and industry partners, with some individuals serving as
external stakeholders, and health clinics or medical facilities as part of their medical centers.
Systems 2024, 12, 215 6 of 29

The review of the existing literature revealed a predominant focus on humanitarian

logistics and supply chains at a macro level, often overlooking the specific needs and
challenges encountered by individual organizations. Consequently, there exists a pressing
need for research that effectively addresses this gap and offers practical insights tailored
to the unique contexts of organizations. Additionally, given the magnitude of disasters
and the delayed assistance from governmental and public institutions in affected regions,
many organizations may experience prolonged periods without aid. Nevertheless, these
organizations possess resources that, if effectively managed, can significantly mitigate the
adverse effects of disasters. Consequently, the proposed topic stands apart from previous
works in its comprehensive consideration of various perspectives, including applicable
areas, centers, and networks. The main contributions of the research are as follows:
• Providing a pre- and post-disaster management model inside an organization using
the system’s internal capabilities and which is independent of other organizations.
• Presenting a mathematical model for pre-and post-disaster planning in an organi-
zation considering operations centers/field offices and external stakeholders as the
relief parts.
• Taking into account the personnel stationed at operations centers/field offices, and
involving external stakeholders to assist in rescue activities and the distributing re-
lief items.
• Taking into account regional branches, major transportation hubs, operational units,
operations centers/field offices, external stakeholders, and medical centers as the
affected areas.
• Considering the regional branches, operations centers/field offices, external stake-
holders, major transportation hubs, and operational units as candidate sites for open-
ing PRCs.
• Considering two types of injuries in the affected areas.
• Taking the total time that injured staffs waited in EMVs or AAs to arrive medical
centers into account.
• Considering different performance levels of the rescue groups of various medical centers.

3. Problem Description and Formulation

In this study, optimal strategies for humanitarian logistics and supply chain man-
agement in organizational contexts are derived. Taking into account that this model is
originally designed for pre-and post-disaster planning in an organization, all internal com-
ponents of the organization are considered, ensuring that it can optimally implement the
required activities for response and preparedness phase. The network consists of affected
areas, PRCs, TRCs, operations centers/field offices, external stakeholders, and medical
centers. Also, the affected areas include regional branches, major transportation hubs, oper-
ational units, operations centers/field offices, external stakeholders, and medical centers. In
the first phase, the PRCs are located and the amount of relief items is pre-positioned in each
established PRC. There are some candidate sites to establish the PRCs. It is assumed that
the PRCs with not-identical capacity can be opened in the regional branches, operations
centers/field offices, external stakeholders, major transportation hubs, operational units in
addition to some proposed sites for PRCs. Furthermore, various relief items such as water,
canned food, and others are strategically positioned in advance, taking into account their
volume and weight. After a disaster occurs and in the second phase, the TRCs’ locations
and donated RIs are specified. Donated RIs are directly sent to TRCs, and the capacity for
all of them is the same. In addition, the number of transferred RIs from TRCs and PRCs
to affected areas would be determined. There are some ground vehicles and helicopters
to distribute RIs among the affected areas, and how many RIs with each transportation
system is sent to affected areas is specified in this phase. Furthermore, there are a number
of injuries in the affected areas to be treated by rescue groups transported from medical
centers to these areas. Some emergency response vehicles (ERVs) and AAs are considered as
transportation options for the rescue groups from medical centers. Based on the number of
Systems 2024, 12, 215 7 of 29

available rescue groups in medical centers, ERVs, and AAs, the number of allocated rescue
groups to each area and the required ERVs and AAs are determined. It is assumed that
each rescue group from each medical center can treat a predetermined number of injuries
that may lead to a number of untreated injurers in each affected area, considering the total
rescue groups allocated to the area. Also, a percentage of injuries in each affected area need
to be transported to medical centers after treatment. Consequently, some ERVs and AAs
are considered which may be able to transport these injuries to medical centers. In this case,
the model determines how many ERVs and AAs are needed and how many injuries from
each affected area must be sent to medical centers. Due to the fact that the organization can
utilize the available staff in the operations centers/field offices, and external stakeholders
to help the rescue activities and distribution of RIs in accelerating the process of these
operations, some staff from these centers are transported to the affected areas. It is assumed
that an operations center/field office or external stakeholder does not need staff from other
operations centers/field offices or external stakeholders to help rescue actions and the
Systems 2024, 12, xdistribution of RIs, due to there being enough staff at these centers. Figure 1 demonstrates
FOR PEER REVIEW
the overview of the entire proposed network.

Figure 1. Schematic view1.of

Figure the proposed
Schematic viewmodel.
of the proposed model.

A mixed integer nonlinear programming (MINLP) model is developed to optimally

3.1. Assumptions
locate and assign PRCs and TRCs to affected areas, and specify the amount of stored RIs in
• The main internal parts of the organization are considered for pre-and post-d
TRCs and PRCs, the amount of distribution of RIs, the number of ground vehicles needed,
planning.
as well as helicopters, EMVs, AAs, and ERNs, the number of injuries and rescue groups
transferred among• The operations
medical centerscenters/field
and affected offices,
areas,regional
and thebranches,
number external stakeholders, m
of transported
staff from operations centers/field offices and external stakeholders to other affected considered
centers, major transportation hubs, and operational units are areas. to
affected areas.
The objective function is to minimize the total costs for pre-and post-disaster planning in
•
addition to penalty Somefor
costs potential
untreatedsites, operations
injuries centers/field
and demand, delayedoffices, regional
total time branches, e
of rescue
stakeholders, medical centers, major transportation hubs,
and distribution of RIs operations, and waiting cost of injured persons in EMVs and AAs. and operational un
takenour
In the following section, into account as the
assumptions willcandidate sites for PRCs.
be mentioned.
• The capacities for TRCs are the same.
3.1. Assumptions• Different capacities are considered for PRCs.
• •
The main internalDistribution
parts ofof the
RIs are conducted are
organization using ground vehicles
considered and helicopters.
for pre-and post-
• The rescue groups are transferred from medical centers to affected areas usin
disaster planning.
and AAs.
• The injured are transported from affected areas to medical centers using ER
AAs.
• The number of staff in the operations centers/field offices and external stakeh
number of available rescue groups in medical centers, number of available all-g
vehicles, helicopters, ERVs, and AAs are known.
• The number of injuries and demands in the affected areas are known.
Systems 2024, 12, 215 8 of 29

• The operations centers/field offices, regional branches, external stakeholders, medical

centers, major transportation hubs, and operational units are considered to be the
affected areas.
• Some potential sites, operations centers/field offices, regional branches, external
stakeholders, medical centers, major transportation hubs, and operational units are
taken into account as the candidate sites for PRCs.
• The capacities for TRCs are the same.
• Different capacities are considered for PRCs.
• Distribution of RIs are conducted using ground vehicles and helicopters.
• The rescue groups are transferred from medical centers to affected areas using ERVs
and AAs.
• The injured are transported from affected areas to medical centers using ERVs and AAs.
• The number of staff in the operations centers/field offices and external stakeholders,
number of available rescue groups in medical centers, number of available all-ground
vehicles, helicopters, ERVs, and AAs are known.
• The number of injuries and demands in the affected areas are known.
• Donated RIs are transported to TRC and are distributed among areas.

3.2. Notations
In this section, the notation utilized in developing the mathematical optimization
model is introduced.

Sets
s Set of external stakeholders
i Set of operational units
a Set of major transportation hubs
e Set of regional branches
h Set of medical centers
g Set of operations centers/field offices
r Set of RIs
p Set of candidate sites for PRCs
l ⊂ { g ∪ e ∪ s ∪ h ∪ a ∪ i} Set of affected areas
f Set of capacity type for PRCs
m Set of candidate zones for TRCs
j∈
Set of candidate zones for PRCs
{ p ∪ g ∪ e ∪ s ∪ h ∪ a ∪ i}
v Set of AAs
z Set of ERV
n Set of ground vehicles
o Set of helicopters
q Set of EMVs
Parameters
PEN r Unmet demand penalty of RI r
CAPG n Capacity of ground vehicle n
CAPH o Capacity of helicopter o
CAPEq Capacity of EMV q
CAPAv Capacity of AA v
CAPRSz Capacity of ERV z
U IP Untreated injury penalty cost
PTR The penalty cost of each unit time the rescue operation is delayed
PTD The penalty cost of each unit time that the distributing RIs operation is delayed
NRM h Number of injuries that rescue group of medical center h can treat h can treat
SCPGrjln Shipping cost of RI r from PRC j to the affected area l by ground vehicle n (Currency unit/(km·kg))
SCPH rjlo Shipping cost of RI r from PRC j to the affected area l by helicopter o (Currency unit/(km·kg))
Shipping cost of RI r from TRC m to the affected area l with ground vehicle n (Currency
SCMGrmln
unit/(km·kg))
Systems 2024, 12, 215 9 of 29

SCMH rmlo Shipping cost of RI r from TRC m to the affected area l with helicopter o (Currency unit/(km·kg))
Shipping cost of a rescue group from the medical center h to affected area l (Currency unit/km) by
SCHEhlz
ERV z
Shipping cost of a rescue group from the medical center h to affected area l (Currency unit/km) by
SCH Ahlv
AA v
Shipping cost of staff from the operations center/field office g to the affected area l (Currency
SCSO gl
unit/km)
SCSEsl Shipping cost of staff from the external stakeholder s to the affected area l (Currency unit/km)
CRM h Operational cost of every rescue group from the medical center h
FEC j f Establishing cost of PRC j with capacity f
PECr The buying cost of RI r
HECrj Holding cost of RI r in PRC j
FCMm Fixed cost of opening TRC m
N Il The number of injuries in affected area l
NSROl The number of staff needed to help rescue operations in the affected area l
NSDOl The number of staff needed to help distributing RIs in the affected area l
TSROl The time span that the rescue activity is delayed because of a lack of staff in the affected area l
The time span that the distribution of RIs’ operation is delayed because of a lack of staff in the
TSRD l
affected area l
DDrl RI r demand in the affected area l
NDDr The number of donated RI r
N AO g Number of accessible staff in the operations center/field office g
N AEs Number of accessible staff in the external stakeholder s
N AG n Number of accessible ground vehicles n
N AH o Number of accessible helicopters o
N AAq Number of accessible EMVs q
N AMMH v Number of accessible AAs v
NREz Number of accessible ERVs z
ENPRT The number of staff in every rescue group
DAPG lj Distance between the affected area l and the PRC j by ground vehicle
DAPH lj Distance between the affected area l and the PRC j by helicopter
DATG lm Distance between the affected area l and the TRC m by ground vehicle
DATH lm Distance between the affected area l and the TRC m by helicopter
DAMElh Distance between the affected area l and the medical center h by ERVs or EMV
DAMH lh Distance between the affected area l and the medical center h by helicopter
DAOlg Distance between the affected area l and the operations center/field office g
DAEls Distance between the affected area l and the external stakeholders s
NRTM h Number of rescue groups in the medical center h
CMI h The capacity of the medical center h to receive injuries from the affected areas
VU r Volume unit of RI r
WU r Weight unit of RI r
VC f Volume capacity of a PRC with capacity type f
VOT TRC volume capacity
The percentage of injuries in the affected area l should be transported to the medical center after
Pl
treating
WC Waiting cost an injured person spent in EMVs or AAs to arrive at the medical center per unit time
MSEq Mean speed of the EMV q
MSAv Mean speed of the AA v
MB1 Budget organization before disaster
MB2 Budget organization after disaster
MP Maximum number of PRCs that can be opened
M A big number
Decision variables
The number of staff transferred from the operations center/field office g to the affected area l to help
NSOR gl
rescue operations
The number of staff transferred from the operations center/field office g to the affected area l to help
NSOD gl
the distribution of RIs within the area
Systems 2024, 12, 215 10 of 29

The number of staff transferred from external stakeholders s to the affected area l to help rescue
NSERsl
operations
The number of staff transferred from external stakeholders s to the affected area l to help the
NSED sl
distribution of RIs within the area
UDrl The number of unmet demands for RI r in the affected area l
NNG n The number of ground vehicles needed n
NNH o The number of helicopters needed o
NNEq The number of EMVs q
NN Av The number of AAs v
NNV z The number of ERV z
Number of rescue groups from the medical center h transferred to the affected area l for treating
NRMARhlz
injuries using an ERV z
Number of rescue groups from the medical center h transferred to the affected area l for treating
NRMAH hlv
injuries using an AA v
NU I l Number of untreated injuries in the affected area l
Xj f 1, if PRC is established in the selected site j with capacity f ; 0, O.W.
Ym 1, if TRC is established in the selected site m; 0, O.W.
N I AMlhv Number of injuries transferred from the affected area l to the medical center h with an AA v
N IDMlhq Number of injuries transferred from the affected area l to the medical center h with an EMV q
SSRl Staff shortage in the affected area l to help rescue operations
SSD l Staff shortage in the affected area l to help in the distribution of RIs
QOPrj Quantity of prepositioned RI r at PRC j
QODrm Quantity of donated RI r stored at TRC m
QTPGrjln Transported quantity of RI r from PRC j to the affected area l with a ground vehicle n
QTPH rjlo Transported quantity of RI r from PRC j to the affected area l with a helicopter o
QTTGrmln Transported quantity of RI r from TRC m to the affected area l with a ground vehicle n
QTTH rmlo Transported quantity of RI r from TRC m to the affected area l with a helicopter o
TTIP Entire time that patients with injuries waited in EMVs or AAs to arrive at medical centers

3.3. Mathematical Model

Here, an MINLP model is derived. The goal of the model is to minimize the total cost
of the proposed model. The model is given as follows:

Min Z = PTR ∑ SSRl TSROl + PTD ∑ SSD l TSRD l + U IP∑ NU I l

l l  l 
+∑ ∑ PEN r UDrl + ∑ ∑ QOPrj PECr + HECrj
r l r j
+∑ ∑ FEC j f X j f + WC TTIP
j f  
+∑ ∑ CRMh ∑ NRMARhlz + ∑ NRMAH hlv
h l z v
+∑ ∑ ∑ SCHEhlz DAMElh NRMARhlz
h l z
+∑ ∑ ∑ SCH Ahlv DAMH lh NRMAH hlv
h l v (1)
+∑ ∑ ∑ ∑ SCPGrjln QTPGrjln WU r DAPG lj
r j l n
+∑ ∑ ∑ ∑ SCPH rjlo QTPH rjlo WU r DAPH lj
r j l o
+∑ ∑ ∑ ∑ SCMGrmln QTTGrmln WU r DATG lm
r m l n
+∑ ∑ ∑ ∑ SCMH rmlo QTTH rmlo WU r DATH lm
r m l o  
+∑ ∑ SCSO gl NSOR gl + NSOD gl DAOlg
g l −{ g∪s}
+∑ ∑ SCSEsl ( NSERsl + NSED sl ) DAEls + ∑ FCMm Ym
s l −{ g∪s} m

The cost function Z intends to minimize the total cost of the considered model that
includes the penalty cost of the delayed total time of rescue and distribution of RI operations,
the untreated injuries penalty cost, the unsatisfied demands penalty cost, the buying cost of
Systems 2024, 12, 215 11 of 29

RIs and hold cost of them at PRCs, the establishing cost of PRCs, the waiting cost of injured
persons in EMVs and AAs, the transportation and operation costs of rescue groups, the
RI transportation costs, and the transportation costs of staff from operations centers/field
offices and external stakeholders to affected areas.
S.t.
   
DAMElh DAMH lh
TTIP = ∑ ∑ ∑ N IDMlhq MSEq
+ ∑ ∑ ∑ N I AMlhv
MSAv
(2)
l h q l h v

Constraint (2) calculates the total time of injured persons spend in EMVs or AAs to
arrive in medical centers.

SSRl = NSROl − ∑ NSOR gl − ∑ NSERsl ∀l − { g ∪ s} (3)

g s

Constraint (3) computes staff shortages in the affected area l to help rescue operations.

SSD l = NSDOl − ∑ NSOD gl − ∑ NSED sl ∀l − { g ∪ s} (4)

g s

Constraint (4) computes staff shortages in the affected area l to help in the distribution
of RIs.
!
∑ ∑ N IDMlhq + ∑ ∑ N I AMlhv = Pl ∑ NRMh ∑ NRMARhlz + ∑ NRMAH hlv ∀l (5)
h q h v h z v

Constraint (5) derives the number of injuries transferred from the affected area l to
medical centers with EMVs and AAs.

∑ ∑ N I AMlhv + ∑ ∑ N IDMlhq ≤ CMI h ∀h (6)

l v l q

Constraint (6) guarantees that transferred injuries from the affected areas to every
medical center is lower than the capacity of each medical center.

∑ ∑ QTPGrjln + ∑ ∑ QTPHrjlo ≤ QOPrj ∀r, j (7)

l n l o

Constraint (7) represents the number of transported RIs from every PRC that is not
higher than the stored RIs.

∑ ∑ QTTGrmln + ∑ ∑ QTTHrmlo ≤ QODrm ∀r, m (8)

l n l o

The above constraint ensures that the amount of transported RIs from every TRC is
not higher than stored RIs.

DDrl − ∑ ∑ QTPGrjln − ∑ ∑ QTPH rjlo− ∑ ∑ QTTGrmln − ∑ ∑ QTTH rmlo = UDrl ∀r, l (9)
j n j o m n m o

This constraint (9) obtains the unmet demands for RI r in the area l.
!
NU I l = N I l − ∑ NRM h ∑ NRMARhlz + ∑ NRMAH hlv ∀l (10)
h z v

Constraint (10) calculates the number of untreated injuries in the area l.

∑ Xj f ≤1 ∀j (11)
f
Systems 2024, 12, 215 12 of 29

Inequality (11) stipulates that only one PRC with a specified capacity could be estab-
lished at most on every potential site.

∑ ∑ Xj f ≤ MP (12)
j f

The aforementioned constraint guarantees the upper limit on the number of PRCs that
can be established.

∑ NSOR gl + ∑ NSOD gl ≤ N AO g ∀g (13)

l −{ g∪s} l −{ g∪s}

∑ NSERsl + ∑ NSED sl ≤ N AEs ∀s (14)

l −{ g∪s} l −{ g∪s}

Constraints (13) and (14) guarantee that transported staffs from operations cen-
ters/field offices and external stakeholders are less than the total accessible staff.

∑ QODrm ≤ NDDr ∀r (15)

m

Constraint (15) ensures that the quantity of each relief item stored in TRCs is lower
than the total amount of donated RIs.

∑ ∑ QOPrj PECr + HECrj + ∑ ∑ FEC j f Xj f ≤ MB1

(16)
r j j f
 
∑ ∑ CRM h ∑ NRMARhlz + ∑ NRMAH hlv
h l z v
+ ∑ ∑ ∑ SCHEhlz DAMElh NRMARhlz
h l z
+ ∑ ∑ ∑ SCH Ahlv DAMH lh NRMAH hlv
h l v
+ ∑ ∑ ∑ ∑ SCPGrjln QTPGrjln WU r DAPG lj
r j l n
+ ∑ ∑ ∑ ∑ SCPH rjlo QTPH rjlo WU r DAPH lj (17)
r j l o
+ ∑ ∑ ∑ ∑ SCMGrmln QTTGrmln WU r DATG lm
r m l n
+ ∑ ∑ ∑ ∑ SCMH rmlo QTTH rmlo WU r DATH lm
r m l o  
+ ∑ ∑ SCSO gl NSOR gl + NSOD gl DAOlg
g l −{ g∪s}
+∑ ∑ SCSEsl ( NSERsl + NSED sl ) DAEls + ∑ FCMm Ym ≤ MB2
b l −{ g∪s} m

Constraints (16) and (17) indicate the budget limitation in the pre-and post-disaster
phase, where the former includes the buying and inventory cost in addition to the estab-
lishment cost, and the latter includes the cost of allocating rescue groups and transferring
them to affected zones that could not cost more than the available budget.
!
1
CAPG n ∑ ∑ ∑ VU r QTPGrjln + ∑ ∑ ∑ VU r QTTGrmln
NNG n ≥ ∀n (18)
r j l r m l

!
1
NNH o ≥
CAPG o ∑ ∑ ∑ VU r QTPHrjlo + ∑ ∑ ∑ VU r QTTHrmlo ∀o (19)
r j l r m l

1
CAPEq ∑ ∑ N IDMlhq
NNEq ≥ ∀q (20)
l h
Systems 2024, 12, 215 13 of 29

!
1
NN Av ≥ ENPRT ∑ ∑ NRMAH hlv + ∑ ∑ N I AMlhv ∀v (21)
CAPAv h l l h

ENPRT
CAPRSz ∑ ∑ NRMARhlz
NNV z ≥ ∀z (22)
h l

The constraints (18)–(22) compute the total number of ground vehicles needed, heli-
copters, EMVs, AAs, and ERVs.

NNG n ≤ N AG n ∀n (23)

NNH o ≤ N AH o ∀o (24)
NNEq ≤ N AAq ∀q (25)

NN Av ≤ N AMMH v ∀v (26)
NNV z ≤ NREz ∀z (27)
Constraints (23)–(27) guarantee that the ground vehicles and helicopters needed cannot
be more than the total accessible ground vehicles and helicopters.

∑ ∑ NRMARhlz + ∑ ∑ NRMAH hlv ≤ NRTMh ∀h (28)

l z l v

Constraint (28) guarantees that the transported rescue groups is lower than total
accessible rescue group in every medical center.

∑ QOPrj VU r ≤ ∑ VC f Xj f ∀j (29)
r f

Inequality (29) specifies that the quantity of pre-positioned relief items in each des-
ignated PRC must not exceed its capacity, and if a PRC is established, it is capable of
storing RIs.
∑ QODrm VU r ≤ VOTYm ∀m (30)
r

Inequality (30) indicates that the quantity of pre-positioned relief items in each desig-
nated TRC must not exceed its capacity, and if a TRC is established, it has the capability to
store relief items.

4. Solution Approach
Metaheuristic algorithms have gained significant traction across various domains
due to their ability to provide robust and near-optimal solutions for complex optimiza-
tion problems [48–52]. Their advantages include flexibility, simplicity, and the capacity
to handle large-scale and nonlinear problems effectively [53–59]. These algorithms have
demonstrated success in various fields such as engineering, logistics, and artificial intel-
ligence [60–66]. Among these, the Grasshopper Optimization Algorithm (GOA) stands
out for its efficiency and effectiveness in solving intricate optimization challenges. In this
paper, we explore the application of GOA to solve a developed model, showcasing its
potential to deliver high-quality solutions in a timely manner. This approach is particularly
beneficial for medium- to large-scale problems, where traditional methods may prove to be
time-consuming and economically unfeasible.

Grasshopper Optimization Algorithm

GOA, introduced by Saremi et al. [67], is a variant of evolutionary algorithms derived
from the Genetic Algorithm, and is inspired by the foraging behavior of grasshoppers in na-
ture. GOA operates as a population-based algorithm with the primary objective of seeking
optimal solutions. Its approach aligns with the natural behavior of grasshoppers, where
Systems 2024, 12, 215 14 of 29

the search space is logically divided into exploration and exploitation phases, mirroring
the soft and continuous movement of nymph grasshoppers for exploitation and the abrupt
movement of adult grasshoppers for exploration [67].
GOA has several advantages for solving single-objective problems. By exploring a
larger portion of the search space, it increases the likelihood of discovering global op-
timal compared to single-solution algorithms. Furthermore, the exchange of informa-
tion within the search space among multiple solutions facilitates rapid progress towards
optimal objectives [68]. Additionally, due to the high rates of repulsion and attraction
among grasshoppers, GOA exhibits remarkable capabilities in both exploitation and ex-
ploration. Consequently, considering its documented benefits and performance in various
studies [69,70], it can be asserted that GOA is effective in addressing real-world problems
characterized by unknown search spaces [71,72]. According to [67,73], the steps of the GOA
are presented in the following equations:

Xi = Si + Gi + Ai (31)

In Equation (31), Xi is the situation of the ith grasshopper, Si denotes the social in-
teraction, Gi indicates the gravity force on the ith grasshopper, and Ai denotes the wind
advection. The social interaction’s component as the main search mechanism is derived as
the following:
N
Si = ∑ s(dij )dˆij (32)
j=1,j̸=i

In Equation (32), dij presents the distance between the ith and the ith grasshopper using
the equation dij = x j − xi . Also, s is a function to specify the strength of social forces
and dˆij = x j − xi /dij . Function s is the main part of the social interaction function, which
denotes the direction of movement of the grasshopper in the group (social force) as follows:
−r
s (r ) = f e l − e −r (33)

In Equation (33), the intensity of attraction with parameter f and the attractive length
scale with parameter l have been demonstrated. s(r ) leads to the creation of repulsion and
attraction forces between the grasshoppers, so changing its parameters has a remarkable
effect on the swarm behaviors. As a result, Saremi, Mirjalili, and Lewis [67] developed the
following model to plan an optimization algorithm:
!
N x − x
ubd − lbd  d j i
d
Xi = c ∑ c s s x j − xi d
dij
+ T̂d (34)
j=1,j̸=i

ubd and lbd are the upper and lower bounds of the dth dimension. The parameters T̂d
and c are used as controller parameters to achieve the target (best solution found so far).
The grasshoppers’ interactions and target pursuit lead to updating the best solution, and
parameter c is the main controller parameter, which is computed by Equation (35):

cmax − cmin
c = cmax − l (35)
L
in which cmax and cmin are regarded as the maximum and minimum values of c, respec-
tively. The value of cmax is 1, and cmin is 0.00001. L and l denote the maximum number of
iterations and the current iteration, in order.

5. Computation Study
5.1. Numerical Experiments
To validate the model developed in this study, a series of small and medium-scale
problems are solved. The mathematical model, formulated as the MINLP model, is solved
Systems 2024, 12, 215 15 of 29

using the BARON solver within GAMS version 25.1.2. The computations were executed
on a computer system equipped with an AMD Ryzen 5 3500U processor clocked at 2.10
GHz and 8 GB of RAM. However, the utilization of the exact method for solving problems,
especially those of a large-scale nature, can be time-consuming and cost-ineffective. To
address this challenge, metaheuristic approaches are commonly employed. Therefore, in
this study, GOA are proposed as a solution method for these conditions. Furthermore, the
efficiency of the proposed metaheuristic algorithm is evaluated by comparing the results
obtained from solving some sample problems of small and medium sizes using GOA with
those obtained from the GAMS solver.
Tables 1 and 2 show the scale of all sample problems and some parameter ranges
utilized in the test problems, respectively. It is important to note that the medium- and
large-sized problems are considered by expanding the number of several sets such as RIs,
operations centers/field offices, regional branches, external stakeholders, etc., rather than
small-size tests.

Table 1. Assumed sets in sample problems.

Problem Scale Sample Problems r g e b h a i p f m n o q v z

Small 1 2 2 1 1 2 1 1 2 2 2 1 2 1 1 2
2 2 3 2 2 3 2 2 2 2 2 2 3 2 2 2
3 3 3 3 2 4 3 2 3 2 3 3 4 2 2 3
Medium 4 4 3 4 4 5 4 4 4 4 4 4 5 4 4 4
5 5 4 5 4 6 4 4 5 4 5 4 6 4 4 4
6 5 5 5 4 7 5 5 5 5 6 5 6 5 5 5
Large 7 8 10 7 6 10 7 6 8 5 10 8 8 7 6 7
8 10 15 12 8 12 8 8 10 7 12 10 10 10 12 8

Table 2. The boundaries of several parameters utilized in the assumed examples.

Parameters Values Parameters Values

WUr (kg) U(0.5, 3) NIl U(30, 150)
NDDr (Unit) U(300, 1500) TSROl (h) U(0.05, 0.2)
HECrj (IRR) U(5, 30) MSEq km U(60, 100)
h 
SCPGrjln (IRR) U(1, 3) MSAv km h
U(220, 250)
SCPHrjlo (IRR) U(3, 6) Pl U(0.05, 0.3)
SCMGrmln (IRR) U(1, 3) SCSOgl (IRR) U(5, 15)
SCMHrmlo (IRR) U(4, 8) SCSEsl (IRR) U(5, 15)
SCHEhlz (IRR) U(40, 60) SCHAhlv (IRR) U(80, 120)

5.2. Results and Model Validation

In this section, the optimal results of first and fourth sample problems are presented as
a small and medium size of the developed model, respectively. Table 3 shows some optimal
decision variables for sample problem 2. Also, Figure 2 demonstrates the optimal allocation
of operations centers/field offices to affected areas to help rescue operations. The number
of staff transported through these areas have also been presented in Table 3 (NSOR gl ). The
objective function value for this sample problem is 27,140,952,496 IRR. As can be seen in
Table 3, the NU I l values are high for regional branches, but the range of this variable is
low for the other sites, and values for operations centers/field offices 1 and 2, the medical
center, and major transportation hub 1 are zero. Moreover, the medical center 1 and major
transportation hub 2 are selected to establish PRCs. Based on Table 3, it is observed that the
number of staff transferred from the operations center/field office g3 to the affected area
e1, and operations center/field office g3 to affected area h2, to assist the distribution of RIs
within the area are 35 and 28, respectively (NSOD g3e1 = 35, NSOD g3h2 = 28). Another
example from this table indicates that the quantity of donated RI r3 stored at TRC m1 would
be 334 (QODr3m1 = 334).
 𝑵𝑺𝑶𝑫𝒈𝒍 𝑵𝑵𝑮𝒏 𝑵𝑵𝑬𝒒
𝑔1. ℎ2 𝑔1. 𝑖1 𝑔1. 𝑖2 𝑔3. 𝑒1 𝑔3. 𝑒2 𝑔3. ℎ1 𝑔3. 𝑎2 𝑛1 𝑛2 𝑞1 𝑞2
10 11 12 20 15 8 14 20 25 3 3
𝑵𝑵𝑯𝒐 𝑵𝑵𝑨𝒗 𝑵𝑵𝑽𝒛 𝑸𝑶𝑫𝒓𝒎 𝒀𝒎
Systems 2024, 12, 215 𝑜1 𝑜2 𝑜3 𝑣1 𝑣2 𝑧1 𝑧2 𝑟3. 𝑚1 𝑚1 𝑻𝑻𝑰𝑷 4.58 16 of 29
15 10 10 4 7 15 18 334 1

Table
As 3. Optimal model
the proposed outcomes of sample17
comprises problem 2.
sets, increasing each set by just one unit signif-
icantly elevates the model’s complexity. Consequently, the number of parameters, varia-
NUIl
bles, and constraints remarkably increase. As a result, and based on Table 1, the forth sam-
g3 e1 e2ple problem
b1 b2 complicated
is more h2 than the second
h3 a2and is identified
one, i1 to bei2 a medium-
31 95 99 6 12 1 8 9 3 6
scale problem. Some results of this problem have been presented in Table 4. It is observed
QOPrj that a lack of staff Xjfin affected areas to helpNSER sl operations (𝑆𝑆𝑅 ) is not NSED
rescue zero insl the me-

r1.a2 dium-scale
5205 h1. f 1 problem, dissimilar
1 to the previous
b1.a1 sample 15problem. In addition, the number
b1.a1 15
r2.h1 of opened
4213 a2. f 2PRCs increased,
1 and one site from four proposed
b2.h3 15 sitesb2.h3
for PRCs was opened. 12
For more clarification in this Table, NSORquantity
gl
of donated RI 𝑟2 stored at TRC 𝑚1, and RI
𝑟4 stored at TRC 𝑚3 are 154 and 87, respectively (𝑄𝑂𝐷 = 154, 𝑄𝑂𝐷 = 87). As
g1.h2 g1.a1 g1.i1 g1.i2 g3.e1 g3.e2 g3.h1 g3.h2 g3.a2 g3.e1
12 10 20another example,
10 35a PRC should 45 be established25 in major
28 transportation
10 hubs 𝑎4
35 with ca-
pacity type 𝑓4 (𝑋 = 1).
NSODgl NNG NNEq
Furthermore, the effect of parameters 𝑁𝐼 and 𝐷𝐷 on nobjective function is investi-
g1.h2 g1.i1 g1.i2 g3.e1
gated in Figuresg3.e2 g3.h1 the performance
3 and 4 to verify g3.a2 ofn1the proposed
n2 q1 expected, q2
model. As
10 11 12when the 20 number 15of injuries or8 demand of RIs 14 in affected
20 25 3 cost also 3
areas increases, total
NNHo NNAv
increases. Observing NNV
that the results QODrm are Y
ofz these test problems predictable
m and reasonable
o1 o2 o3demonstrates
v1 that
v2 the model is z1 working correctly
z2 and
r3.m1is verified
m1 to be implemented
TTIP in4.58
15 10 10various size
4 of the7 model. 15 18 334 1

Figure 2. The optimal allocation of operations centers/field offices to affected areas to help rescue
operations (sample problem 2).

As the proposed model comprises 17 sets, increasing each set by just one unit signifi-
cantly elevates the model’s complexity. Consequently, the number of parameters, variables,
and constraints remarkably increase. As a result, and based on Table 1, the forth sample
problem is more complicated than the second one, and is identified to be a medium-scale
problem. Some results of this problem have been presented in Table 4. It is observed that a
lack of staff in affected areas to help rescue operations (SSRl ) is not zero in the medium-
scale problem, dissimilar to the previous sample problem. In addition, the number of
opened PRCs increased, and one site from four proposed sites for PRCs was opened. For
more clarification in this table, quantity of donated RI r2 stored at TRC m1, and RI r4 stored
at TRC m3 are 154 and 87, respectively (QODr2m1 = 154, QODr4m3 = 87). As another
example, a PRC should be established in major transportation hubs a4 with capacity type
f 4 (Xa4 f 4 = 1).
Systems 2024, 12, 215 17 of 29

Table 4. Some results of sample problem 4.

Systems
Systems 2024,
2024, 12,
12, xx FOR
FOR PEER
PEER REVIEW
REVIEW SSRl 19
19 of
of 31
31
e1 e2 e3 e4 h1 h2 h3 h4 h5 i1 i3
55 65 42 52 35 40 25 20 30 50 25
QODrm Xjf
Figure 2. The optimal allocation of operations Ymto help rescue
centers/field offices to affected areas
r2.m1 154 r4.m1 operations
180 (sample
e3. f 4 problem
1 2). a4. f 4 1 p2. f 2 1 m1 1
r2.m3 247 r4.m3 87 b4. f 4 1 i1. f 2 1 m3 1
Table 4. Some results of sample problem 4.

Furthermore, the effect of parameters𝑺𝑺𝑹 N I l𝒍𝒍and DDrl on objective function is investi-

gated in Figures 3 and 4 to verify the performanceℎ2
𝑒1 𝑒2 𝑒3 𝑒4 ℎ1 ℎ3
of the proposedℎ4model.
ℎ5 As 𝑖1 𝑖3
expected,
55 65 42 52 35 40 25 20 30 50
when the number of injuries or demand of RIs in affected areas increases, total cost also 25
𝑸𝑶𝑫𝒓𝒎
increases. Observing
𝒓𝒎 that the results of these test𝑿problems
𝒋𝒇
𝒋𝒇 𝒀𝒎
are predictable and reasonable
𝒎
𝑟2. 𝑚1 154 that
demonstrates 𝑟4.the
𝑚1model
180 𝑒3.is 𝑓4 𝑎4. 𝑓4 and
1 correctly
working 1 is 𝑝2. 𝑓2 1to be implemented
verified 𝑚1 1 in
𝑟2. 𝑚3 size
various 247 𝑟4.model.
of the 𝑚3 87 𝑏4. 𝑓4 1 𝑖1. 𝑓2 1 𝑚3 1

Figure3.3.Objective
Figure Objective function vs. N𝑁𝐼
functionvs. I l ..

Figure4.4.Objective
Figure Objective function
functionvs. 𝐷𝐷rl ..
vs. DD

5.3. GOA Parameters Tuning

5.3. GOA Parameters Tuning
In order to use meta-heuristic algorithms, it is essential to modify their parameters
In order to use meta-heuristic algorithms, it is essential to modify their parameters
for obtaining acceptable solutions. In this section, GOA parameters such as grasshopper’s
for obtaining acceptable solutions. In this section, GOA parameters such as grasshopper’s
population scale (NPOP) and iterations number (NI) are determined using the Taguchi
population scale (NPOP) and iterations number (NI) are determined using the Taguchi
method by MINITAB. These parameters and their different level are demonstrated in
method by MINITAB. These parameters and their different level are demonstrated in Ta-
Table 5. The L27(32) experiment was applied to adjust the parameters of GOA. After
ble 5. The L27(32)
conducting experiment
experiments was applied
in MINITAB, to adjustmean
corresponded the parameters of GOA.ratio
and signal-to-noise After con-
(SNR)
ductingwere
graphs experiments
obtained, in
as MINITAB, corresponded
shown in Figures 5 and 6, mean and signal-to-noise ratio (SNR)
respectively.
graphs were obtained, as shown in Figures 5 and 6, respectively.
According to the mean graphs (Figure 5), the minimum mean was at the second level
for NPOP and the third level for NI. In the SNR graph (Figure 6), the maximum signal-to-
noise was at the second and third level for the parameters, respectively. Consequently, the
optimal values for NPOP and NI were 80 and 120, respectively.
Systems 2024, 12, 215 18 of 29
Systems 2024, 12, x FOR PEER REVIEW 20 of 31
Systems 2024, 12, x FOR PEER REVIEW 20 of 31

Table 5. The level of GOA parameters.

Table 5. The level of GOA parameters.
Table 5. The level of GOA parameters. Factors
Level Factors
Level NPOP Factors NI
Level NPOP NI
1 NPOP 60 NI 80
1 60 80
1 2 60 80 100
80
2 3 80 100 100
120
2 80 100
3 100 120
3 100 120

Main Effects Plot for Means

Main Effects
Data Plot
Meansfor Means
Data Means
NPOP NI
2.8300E+11 NPOP NI
2.8300E+11
2.8200E+11
2.8200E+11
2.8100E+11
2.8100E+11
of Means

2.8000E+11
of Means

2.8000E+11
2.7900E+11
2.7900E+11
Mean

2.7800E+11
Mean

2.7800E+11
2.7700E+11
2.7700E+11
2.7600E+11
2.7600E+11
2.7500E+11
2.7500E+11
2.7400E+11
2.7400E+11 1 2 3 1 2 3
1 2 3 1 2 3

Figure5.5.Mean
Figure Meangraph
graphof
ofTaguchi
Taguchiexperiment.
experiment.
Figure 5. Mean graph of Taguchi experiment.

Main Effects Plot for SN ratios

Main Effects Plot
Data for SN ratios
Means
Data Means
NPOP NI
NPOP NI
-228.80
-228.80
ratios

-228.85
ratios

-228.85
of SN

-228.90
of SN

-228.90
Mean

-228.95
Mean

-228.95

-229.00
-229.00

-229.05
-229.05 1 2 3 1 2 3
1 2 3 1 2 3
Signal-to-noise: Smaller is better
Signal-to-noise: Smaller is better

Figure6.6.SNR
SNR graphof
of Taguchiexperiment.
experiment.
Figure 6. SNRgraph
Figure graph ofTaguchi
Taguchi experiment.
5.4. Comparative
According toExperiments
the mean graphs (Figure 5), the minimum mean was at the second level
5.4. Comparative Experiments
for NPOP and the third level
Table 6 represents for NI. In the
a comparative SNR graph
analysis for six(Figure
sample6),teststhe presented
maximum in signal-to-
Table 1
noise Table
wasGAMS6 represents
at the second a comparative
and Including analysis
third levelnumerous for six
for the parameters, sample tests presented in Table
the1
between and GOA. sets in therespectively.
proposed model Consequently,
substantially
between
optimal GAMS
values and GOA.
for NPOP time Including
and NI were numerous sets
80 andproblems in the
120, respectively. proposed model substantially
increases computational when solving with exact methods, particularly as
increases computational time when solving problems with exact methods, particularly as
the indices rise. Therefore, employing an exact approach for large-scale problems is nei-
the Comparative
5.4. indices rise.Experiments
Therefore, employing an exact approach for large-scale problems is nei-
ther rational nor economical. In such scenarios, resorting to a metaheuristic algorithm pre-
ther Table
rational nor economical.
6 represents In such scenarios,
a comparative analysis resorting to a metaheuristic
for sixefficiently.
sample tests algorithm
presented in how pre-
Table 1
sents a more suitable solution for tackling the model Figure 7 shows so-
sents a more
between GAMS suitable
and solution
GOA. for tackling
Including the model
numerous sets inefficiently.
the proposed Figure 7 shows
model how so-
substantially
lution time exponentially rises as the scale of problem rises.
lution time
increases exponentially
computational rises as the scale of problemwith rises.exact methods, particularly as
To demonstrate thetime when
efficiency solving
of problems
the proposed meta-heuristic algorithm, the gap rate
the To demonstrate
indicesGAMS the
rise. Therefore, efficiency of the proposed meta-heuristic algorithm, the gap rate
between and GOAemploying
is derived an exact
using approach
Equation (36)foraslarge-scale
follows: problems is neither
between GAMS and GOA is derived using Equation (36) as follows:
rational nor economical. In such scenarios, resorting to a metaheuristic algorithm presents
a more suitable solution for tackling 100 ∗ (𝑍
the=model − 𝑍 Figure
efficiently. ) 7 shows how solution
𝐺𝑎𝑝 𝑟𝑎𝑡𝑒 100 ∗ (𝑍 −𝑍 ) (36)
time exponentially rises as the 𝐺𝑎𝑝 scale𝑟𝑎𝑡𝑒 =
of problem 𝑍
rises. (36)
𝑍
in which 𝑍 and 𝑍 are the optimal cost of GAMS and GOA, respectively. Accord-
in which 𝑍 and 𝑍 are the optimal cost of GAMS and GOA, respectively. Accord-
ing to the obtained gaps in Table 6, There is negligible gap observed between the two
ing to the obtained gaps in Table 6, There is negligible gap observed between the two
solvers in terms of the cost function value. Also, Table 6 shows that the computational time
solvers in terms of the cost function value. Also, Table 6 shows that the computational time
 of GOA is remarkably lower than exact approach as the scale of problems increases. There-
fore, it can be concluded that the GOA is reliable and qualified, and can be applied to
large-scale problems. Figure 8 depicts the convergence graph for sample problem 4. Fi-
nally, the results of sample problems 7 and 8 as two large-scale problems are presented in
Systems 2024, 12, 215 Table 6. As observed, the GOA can solve the large-scale problems in an acceptable and
19 of 29
economical time.

Table 6. Solutions obtained using GOA and GAMS for sample problems.
Table 6. Solutions obtained using GOA and GAMS for sample problems.
Value of Objective Functions CPU Engagement Time
Sample Problems Value of Objective Functions CPU Engagement Time Gap%
Sample Problems GAMS GOA GAMS GOA Gap%
GAMS GOA GAMS GOA
Systems 2024, 12, x FOR PEER REVIEW 1 5,035,880,177 5,036,745,935 4.1″ 40.2″ 21 of 0.0131
12 5,035,880,177 5,036,745,935 ′′ ′′
27,140,952,496 27,283,427,521 4.1 ′′ 12.5″ 40.246.7″ 0.01
0.52
2 27,140,952,496 27,283,427,521 12.5 46.7′′ 0.52
3 32,750,214,458 33,028,439,877 120.2″ 50.3″ 0.84
3 32,750,214,458 33,028,439,877 120.2′′ 50.3′′ 0.84
4
of GOA is remarkably
4 lower43,948,560,292 44,513,486,111
than exact44,513,486,111
43,948,560,292 approach as the scale 500.1″
′′
500.1of problems70.5 70.5″
′′ 1.28
increases. There-
1.28
5 concluded 49,799,224,339 ′′ ′′
fore, it can 5be 49,799,224,339
that the GOA is50,780,269,058
50,780,269,058
reliable and1300.3 1300.3″
qualified, can77.4″
and 77.4 1.96
1.96
be applied to
′′ ′′
large-scale 6problems.
6 58,218,472,604
Figure 8 depicts59,621,537,793
58,218,472,604 convergence2700.9
the59,621,537,793 graph 2700.9″ 84.984.9″
for sample problem 2.40
2.40
4. Fi-
7 - 101,342,146,208 - 115.3′′ -
7 of sample problems
nally, the results - 7 and101,342,146,208 -
8 as two large-scale problems 115.3″
are presented -in
8 - 141,075,529,112 - 149.4′′ -
8
Table 6. As observed, the GOA -can solve141,075,529,112
the large-scale problems - in an 149.4″
acceptable and -
economical time.

Table 6. Solutions obtained using GOA and GAMS for sample problems.

Value of Objective Functions CPU Engagement Time

Sample Problems Gap%
GAMS GOA GAMS GOA
1 5,035,880,177 5,036,745,935 4.1″ 40.2″ 0.01
2 27,140,952,496 27,283,427,521 12.5″ 46.7″ 0.52
3 32,750,214,458 33,028,439,877 120.2″ 50.3″ 0.84
4 43,948,560,292 44,513,486,111 500.1″ 70.5″ 1.28
5 49,799,224,339 50,780,269,058 1300.3″ 77.4″ 1.96
6 58,218,472,604 59,621,537,793 2700.9″ 84.9″ 2.40
7 - 101,342,146,208 - 115.3″ -
Figure7.7.CPU
Figure CPUtime
timefor
forproposed
proposedsample
sampleproblems.
problems.
8 - 141,075,529,112 - 149.4″ -
To demonstrate the efficiency of the proposed meta-heuristic algorithm, the gap rate
between GAMS and GOA is derived using Equation (36) as follows:

100 ∗ ( ZGOA − ZGAMS )

Gap rate = (36)
ZGAMS

in which ZGAMS and ZGOA are the optimal cost of GAMS and GOA, respectively. According
to the obtained gaps in Table 6, There is negligible gap observed between the two solvers in
terms of the cost function value. Also, Table 6 shows that the computational time of GOA is
remarkably lower than exact approach as the scale of problems increases. Therefore, it can be
concluded that the GOA is reliable and qualified, and can be applied to large-scale problems.
Figure 8 depicts the convergence graph for sample problem 4. Finally, the results of sample
problems 7 and 8 as two
Figure 8. Convergence large-scale
of GOA problems
to the optimal are for
solution presented in Table4.6. As observed, the
sample problem
Figure
GOA7.canCPU timethe
solve for large-scale
proposed sample problems.
problems in an acceptable and economical time.
6. Sensitivity Analysis
Conducting sensitivity analysis helps to identify the primary influential parameters
and elucidate their impact on optimal solutions and the objective function. This analysis
facilitates a comprehensive comprehension of the model’s behavior. The results of the sen-
sitivity analysis are derived in Figures 9–18. Figures 9 and 10 show the impact of the avail-
able budget before disaster (𝑀𝐵 ) on objective function and total unsatisfied demands of
affected areas, respectively. As can be seen in Figure 9, when the budget before disaster

Figure
Figure 8. 8. Convergence
Convergence ofof
GOAGOA
toto the
the optimal
optimal solution
solution forfor sample
sample problem
problem 4.4.

6. Sensitivity Analysis
Conducting sensitivity analysis helps to identify the primary influential parameters
and elucidate their impact on optimal solutions and the objective function. This analysis
facilitates a comprehensive comprehension of the model’s behavior. The results of the sen-
sitivity analysis are derived in Figures 9–18. Figures 9 and 10 show the impact of the avail-
Systems 2024, 12, 215 20 of 29

6. Sensitivity Analysis
Conducting sensitivity analysis helps to identify the primary influential parameters
and elucidate their impact on optimal solutions and the objective function. This analysis
Systems 2024, 12, x FOR PEER REVIEW
facilitates a comprehensive comprehension of the model’s behavior. The results 22 of ofthe31
Systems 2024, 12, x FOR PEER REVIEW 22 of 31
sensitivity analysis are derived in Figures 9–18. Figures 9 and 10 show the impact of the
available budget before disaster (MB1 ) on objective function and total unsatisfied demands
of affected by
decreases areas,
10%respectively.
and 20%, theAs cancost
total be seen in Figure
increases about 9, when
0.5% and the 2.5%,
budget before disaster
respectively, and
decreases
decreases by 10% and 20%, the total cost increases about 0.5%
as seen in Figure 10, the unsatisfied demands also increases. This is becauserespectively,
by 10% and 20%, the total cost increases about 0.5% and 2.5%,
and respectively,
2.5%, either some and
as
and
PRCsseen in
ashave Figure
seenbeen 10,
in Figure
closed,the unsatisfied
10,orthe
PRCsunsatisfied demands
with lower demandsalso
capacities increases.
also increases.
have This
been opened.is because
This isOn either
because some
either
the opposite
PRCswhen
some
side, have been
PRCs 𝑀𝐵 closed,
have been or PRCs
closed,
increases by or with
10%,PRCs lower
with capacities
unsatisfied lower have
capacities
demands beenhave
considerablyopened. On theand
beendecrease
opened. opposite
Ontotal
the
side,increases
opposite
cost when side, 𝑀𝐵dueincreases
when toMB by 10%,
increases
the1opening newunsatisfied
by anddemands
10%, unsatisfied
PRCs increasing considerably
demands decrease
considerably
distribution and after
total
decrease
of RIs costs
costtotal
and increases
disaster. due toasthe
cost increases
Moreover, 𝑀𝐵
dueopening
to the new
increasesopeningPRCs
by new
20%, and increasing
PRCs
unsatisfied distribution
anddemands
increasing and of RIs
distribution
the totalcostsof after
cost RIs
de-
disaster.
costs after Moreover,
disaster. as 𝑀𝐵
Moreover,
crease. This may be because increasing increases
as MB 1 by 20%,
increases unsatisfied
by 20%, demands
unsatisfied and
demands
the capacity of PRCs can be less costly than the the total
and cost
the de-
total
crease.
cost This
decrease. may
This be
maybecause
be increasing
because the
increasing capacity
the of
capacity PRCs
of
penalty for unsatisfied demands, which has decreased once again. Figures 11 and 12 show can
PRCs be
can less
be costly
less than
costly thanthe
penalty
the penaltyfor unsatisfied
for unsatisfied demands,
demands, which has
which decreased
has decreasedonce again.
once
the impact of the available budget after disaster (𝑀𝐵 ) on objective function and total un- Figures
again. 11
Figures and 11 12
andshow12
the impact
show
satisfied of theof
the demands
impact available
theaffected
of budget
available after
budget
areas, disaster (𝑀𝐵It is) on
after disaster
respectively. (MB objective
observed2 ) on that function
objective and
the total
whenfunction 𝑀𝐵 and un-
in-
satisfied
total demands
unsatisfied of
demands affected
of areas,
affected respectively.
areas, It
respectively. is observed
It is observed
creases, total cost and unsatisfied demands are reduced. Considering this point that the that when
that the
when 𝑀𝐵
the MB in-
2
increases,
creases, total
total cost
cost and
and unsatisfied
unsatisfied demands
demands are reduced. Considering
fixed cost for opening TRCs is significantly lower than PRCs, increasing the budget after this
this point
point that
that the
the
fixed
fixedcost
disastercost for
has opening
formore
opening
impactTRCs
TRCsonistotal
issignificantly
cost ratherlower
significantly than than
lower thanPRCs,
budget PRCs,
before increasing
disaster.the
increasing thebudget
budget
Another after
after
reason
disaster
disaster has
has more
more impact
impact onontotal
totalcost
costrather
ratherthan
thanbudget
budget before
before
is that costs associated after disaster include many components such as the cost of distri-disaster.
disaster. Another
Another reason
reason is
that costs
bution associated
is that costs
of RIs,associated after disaster
after disaster
the transportation include many
include
of rescue components
many staff.,
groups, components such as
etc., allsuch the cost of
as theare
of which distribution
cost of distri-
influenced
of RIs,
bution
by 𝑀𝐵 . the
of transportation
RIs, the of rescue
transportation groups,
of rescue staff.,
groups,etc., all
staff.,of which
etc., all are
of influenced
which are by MB2 .
influenced
by 𝑀𝐵 .

Figure9.9.Objective
Figure Objective function
functionvs. 𝑀𝐵 ..
vs. MB
Figure 9. Objective function vs. 𝑀𝐵1 .

Figure 10. Unsatisfied demands (𝑈𝐷 ) vs. 𝑀𝐵 .

Figure 10.Unsatisfied
Figure10. Unsatisfieddemands
demands(UD vs. MB
(𝑈𝐷rl )) vs. 𝑀𝐵1 ..
 Systems2024,
Systems 2024,12,
12,215
x FOR PEER REVIEW 23
21 of 31
of 29

Figure 11. Objective function vs. 𝑀𝐵 .

Figure 11. Objective

Figure11. Objective function vs. MB
function vs. 𝑀𝐵2 ..

Figure 12. Unsatisfied demands (𝑈𝐷 ) vs. 𝑀𝐵 .

Figures 13 and 14 show the impact of mean speed of AA on the entire time that pa-
tients with injuries spend in EMVs or AAs to arrive medical centers (𝑇𝑇𝐼𝑃) and objective
functions. As observed, providing that the speed of AAs increase, 𝑇𝑇𝐼𝑃 can considerably
decrease, which resulted in saving the lives of patients with injuries. Therefore, even a five
percent increase in speed has a significant impact on the efficiency of the system. Also, the
total cost is affected by changing speed. Using the newest and advanced AAs in the or-
Figure
Figure 12. Unsatisfied
12.
ganization can avoiddemands
Unsatisfied demands (UD
(𝑈𝐷lives
losing the vs.ofMB
rl )) vs. the2.. injured
𝑀𝐵

Figures 13 and 14 show the impact of mean speed of AA on the entire time that pa-
tients with injuries spend in EMVs or AAs to arrive medical centers (𝑇𝑇𝐼𝑃) and objective
functions. As observed, providing that the speed of AAs increase, 𝑇𝑇𝐼𝑃 can considerably
decrease, which resulted in saving the lives of patients with injuries. Therefore, even a five
percent increase in speed has a significant impact on the efficiency of the system. Also, the
total cost is affected by changing speed. Using the newest and advanced AAs in the or-
ganization can avoid losing the lives of the injured

Figure 13.TTIP
Figure13. vs. MSA
TTIPvs. 𝑀𝑆𝐴v ..
Figures 13 and 14 show the impact of mean speed of AA on the entire time that
patients with injuries spend in EMVs or AAs to arrive medical centers (TTIP) and objective
functions. As observed, providing that the speed of AAs increase, TTIP can considerably
decrease, which resulted in saving the lives of patients with injuries. Therefore, even a five
percent increase in speed has a significant impact on the efficiency of the system. Also,
the total cost is affected by changing speed. Using the newest and advanced AAs in the
organization can avoid losing the lives of the injured
Figure 13. TTIP vs. 𝑀𝑆𝐴 .
Figures 15 and 16 indicate the impact of the number of available EMVs (N AAq )
on TTIP, objective function, and NU I l . As can be seen, when N AAq increases, TTIP
increases, but the total costs and number of untreated injured decreases. It is because the
number of injuries transported from the affected areas to medical centers with EMVs and
AAs increases and decreases, respectively, by increasing N AAq . Hence, when more AAs
are used to transport rescue groups from medical centers to affected areas, the number
Systems 2024, 12, 215 22 of 29

of untreated injured decreases, as shown in Figure 16. The impact of the number of
available AAs (N AMMH v ) on TTIP, objective function, and NU I l is evaluated through
Figures 17 and 18. As observed, by increasing total available AAs, TTIP decreases, and the
total costs and number of untreated injured decrease. This is because the number of injuries
transported from affected areas to medical centers decreases with EMVs and increases with
Systems 2024, 12, x FOR PEER REVIEW 24 of 31
AAs, resulting in a decrease in TTIP. Also, the total number of untreated injured decreases,
and hence, total cost decreases.
Figure 14. Objective function vs. 𝑀𝑆𝐴 .

Figures 15 and 16 indicate the impact of the number of available EMVs (𝑁𝐴𝐴 ) on
𝑇𝑇𝐼𝑃 , objective function, and 𝑁𝑈𝐼 . As can be seen, when 𝑁𝐴𝐴 increases, 𝑇𝑇𝐼𝑃 in-
creases, but the total costs and number of untreated injured decreases. It is because the
number of injuries transported from the affected areas to medical centers with EMVs and
AAs increases and decreases, respectively, by increasing 𝑁𝐴𝐴 . Hence, when more AAs
are used to transport rescue groups from medical centers to affected areas, the number of
untreated injured decreases, as shown in Figure 16. The impact of the number of available
AAs (𝑁𝐴𝑀𝑀𝐻 ) on 𝑇𝑇𝐼𝑃, objective function, and 𝑁𝑈𝐼 is evaluated through Figures 17
and 18. As observed, by increasing total available AAs, 𝑇𝑇𝐼𝑃 decreases, and the total
costs and number of untreated injured decrease. This is because the number of injuries
transported from affected areas to medical centers decreases with EMVs and increases
with AAs, resulting in a decrease in 𝑇𝑇𝐼𝑃. Also, the total number of untreated injured
Figure 14.Objective
Figure14. function vs. MSA v ..
decreases, Objective
and hence,function cost𝑀𝑆𝐴
total vs. decreases.
Figures 15 and 16 indicate the impact of the number of available EMVs (𝑁𝐴𝐴 ) on
𝑇𝑇𝐼𝑃 , objective function, and 𝑁𝑈𝐼 . As can be seen, when 𝑁𝐴𝐴 increases, 𝑇𝑇𝐼𝑃 in-
creases, but the total costs and number of untreated injured decreases. It is because the
number of injuries transported from the affected areas to medical centers with EMVs and
AAs increases and decreases, respectively, by increasing 𝑁𝐴𝐴 . Hence, when more AAs
are used to transport rescue groups from medical centers to affected areas, the number of
untreated injured decreases, as shown in Figure 16. The impact of the number of available
AAs (𝑁𝐴𝑀𝑀𝐻 ) on 𝑇𝑇𝐼𝑃, objective function, and 𝑁𝑈𝐼 is evaluated through Figures 17
and 18. As observed, by increasing total available AAs, 𝑇𝑇𝐼𝑃 decreases, and the total
costs and number of untreated injured decrease. This is because the number of injuries
transported from affected areas to medical centers decreases with EMVs and increases
with AAs, resulting in a decrease in 𝑇𝑇𝐼𝑃. Also, the total number of untreated injured
decreases,
Figure andvs.
15. TTIP hence,
N AAtotal
q.
cost decreases.
Figure 15. TTIP vs. 𝑁𝐴𝐴 .

Figure16.
Figure TTIP vs. 𝑁𝐴𝐴
15.Objective . and NU I vs. N AA .
function l q
Systems 2024,
Systems 2024, 12,
12, xx FOR
FOR PEER
PEER REVIEW
REVIEW 25 of
25 of 31
31

Systems 2024, 12, 215 23 of 29

Figure 16.
Figure 16. Objective
Objective function
function and 𝑁𝑈𝐼 vs.
and 𝑁𝑈𝐼 𝑁𝐴𝐴 ..
vs. 𝑁𝐴𝐴

Figure
Figure 17.TTIP
Figure17.
17. TTIP vs.N𝑁𝐴𝑀𝑀𝐻
TTIPvs.
vs.
AMMH v. .
𝑁𝐴𝑀𝑀𝐻 .

Figure 18.Objective
Figure18.
Figure 18. Objectivefunction
Objective functionand
function and 𝑁𝑈𝐼
andNU I l vs.
𝑁𝑈𝐼 vs. 𝑁𝐴𝑀𝑀𝐻
vs. N AMMH.v. .
𝑁𝐴𝑀𝑀𝐻
7. Discussions and Managerial Implications
7. Discussions
7. Discussions and and Managerial
Managerial Implications
Implications
This study pioneers the development of a mathematical optimization model tailored
This study
This study pioneers
pioneers the the development
development of of aa mathematical
mathematical optimization
optimization model model tailored
tailored
specifically for an organization-based humanitarian logistics and supply chain network,
specifically for
specifically for an
an organization-based
organization-based humanitarian
humanitarian logistics logistics andand supply
supply chain
chain network,
network,
encompassing pre- and post-disaster planning, a novel endeavor not previously explored.
encompassing pre-
encompassing pre- and post-disaster
post-disaster planning, aa novel novel endeavor not not previously
previously explored.
explored.
In contrast to prior and research focusingplanning, on humanitarianendeavor relief operations, this study rep-
In contrast
In contrast to prior research
to prior research focusing on humanitarian
focusingbyonuniquely
humanitarian relief operations,
relief all
operations, this study
thisorganization,repre-
study repre-
resents a groundbreaking approach integrating facets of an
sents a groundbreaking
sents a groundbreaking approach by
approachoffices, uniquely
by uniquely integrating
integrating all facets
all facets of an organization,
of centers,
an organization,
including operations centers/field external stakeholders, medical regional
including operations
including operations centers/field
centers/field offices,
offices, external
external stakeholders,
stakeholders, medicalmedical centers,
centers, regional
regional
branches, major transportation hubs, and operational units. In other words, the litera-
branches,
branches, major
major transportation
transportation hubs,
hubs, and
and operational
operational units.
units.
ture review reveals that most existing research in this field have studied humanitarian In
In other
other words,
words, the
the literature
literature
review reveals
review
logistics reveals
and supplythat most
that most existing
existing
chains outside research
research in this
in
the confines this field
field have studied
of a have
single studied humanitarian
humanitarian
organization. logistics
These logistics
studies
and supply
and supply
typically chains
chains
involve outside the confines
outside the confines
non-governmental of a single
of a single(NGOs),
organizations organization.
organization. These studies
These studies
governmental typically
typically
organizations
involvethe
involve
(GOs), non-governmental
non-governmental
Red Cross, and other organizations
organizations (NGOs),
relevant(NGOs), governmental
governmental
entities conducting organizations
organizations
humanitarian (GOs), the
(GOs),
activities the
to
Red
Red Cross,
Cross, and
and other
other relevant
relevant entities
entities conducting
conducting humanitarian
humanitarian
manage disasters in broader regions, cities, or areas. To our knowledge, no prior study has activities
activities to
to manage
manage dis-
dis-
asters in
asters
presentedin broader
broader regions, cities,
regions,
a comprehensive cities, or areas.
or areas. To
mathematical To our knowledge,
our knowledge,
optimization model nofor
no prior
prior study has
study
disaster has presented
presented
management
a
a comprehensive
comprehensive mathematical
mathematical optimization
optimization model
model for
for disaster
disaster
within an organization that concurrently addresses all aspects, including the distribution management
management within an
within an
organization
organization
of that concurrently
that concurrently
RIs, the transfer of rescue groups addresses
addresses all aspects,
all aspects,
from medical including
including
centers the distribution
the distribution
to affected of RIs,
of RIs, the
areas, the transporta- the
transfer
transfer
tion of rescue
of rescue injuries
of emergency groups from
groups from
to medical
medical
medical centers
centers
centers, thetotomobilization
affected areas,
affected areas, the transportation
the
of staff transportation
from operations of
of
emergency injuries
emergency
centers/field injuries
offices and to medical
to medical centers,
external centers, the mobilization
the
stakeholders mobilization
to affected areas of staff
of staff from operations
from
for distributionoperations
of RIs andcen-
cen-
ters/field
ters/field
rescue officessupport,
offices
activities and external
and external
and thestakeholders
stakeholders
utilizationto toofaffected
affected areas for
areas
all transportation for distribution
distribution of RIs
of
vehicles available RIs and
and res-
res-
within
cue
cue activities
activities support,
support, and
and the
the utilization
utilization of
of all
all transportation
transportation
the organization. Moreover, the core of this study is a novel approach to humanitarian vehicles
vehicles available
available within
within
the organization.
the organization.
logistics and supply Moreover,
Moreover, the core
the
chain management, core of
of focusing
this study
this study onisismanaging
aa novel
novel approach
approach
disastersto to humanitarian
humanitarian
within an orga-
logistics
nization and
logistics without supply
and supply chain management,
chain assistance,
external management, focusing
an focusing
area thaton on managing
hasmanaging disasters
not beendisasters
previously within
within an organ-
an
explored. organ-
By
ization without
utilizing
ization without
the proposedexternal
external assistance,
model, an area
area that
an organization
assistance, an that
can has
has not been
effectively
not been and
previously explored.
independently
previously explored. By uti-
manage
By uti-
lizing the
disasters
lizing the proposed
within
proposed model,operations.
its internal
model, an organization
an organization can can effectively
effectively and and independently
independently manage manage
disasters
disasters withinaits
Specifically,
within its internal
notable
internal operations.
distinction
operations. in the proposed model is the inclusion of the waiting
time for injured individuals in EMVs or AAs until they reach medical centers—a factor
Systems 2024, 12, 215 24 of 29

overlooked in prior research. This waiting time is incorporated as a critical criterion within
the objective function of this study. The reason is that it is vital for an organization to
quickly attend to its human resources after a disaster occurs. Another significant departure
in the theoretical framework of this research, as opposed to previous studies, is the acknowl-
edgment of two vital organizational components as relief centers. These centers, namely
operations centers/field offices and external stakeholders, are recognized as integral relief
hubs within the organization. This recognition marks a departure from prior works, which
typically focused on external entities for relief efforts. For example, operations centers/field
offices are considered to be relief centers within an organization because they typically have
their own resources and personnel to support their functions and operations. They include
staff such as operations managers, field workers, coordinators, and support teams who
collaborate to ensure the smooth operation of the organization’s activities. Additionally,
they often have access to local resources and partnerships, which enable them to effectively
address local needs and facilitate their activities.
Regarding the general applicability of the proposed model to different organizations, it
is important to note that the components were selected based on their broad relevance and
critical roles in humanitarian logistics and supply chain management within organizations.
The following are the reasons why these components are broadly applicable:
• Regional Branches: These act as localized hubs that can quickly respond to regional
needs. Most organizations, regardless of their industry, have a decentralized structure
with regional branches to ensure efficient operations and responsiveness. That is why
these centers can be considered to be candidate zones for PRCs.
• Operational Units: These units are the backbone of an organization’s response mecha-
nism, handling everything from logistics to administration. Their presence is universal
across organizations to ensure operational continuity and efficiency. That is why these
centers can be considered as candidate zones for PRCs.
• Major Transportation Hubs: Effective disaster management and humanitarian logistics
depend on the ability to quickly move resources. Major transportation hubs are crucial
for facilitating the rapid distribution of supplies and personnel, making them a vital
component in any organization’s logistics network.
• Operational Centers/Field Offices: These centers are pivotal for coordinating on-
ground activities and managing logistics. Almost all organizations have some form of
operational centers or field offices to oversee their day-to-day activities and emergency
responses. This clarification confirms why these centers can be considered to be relief
centers.
• External Stakeholders: Collaborations with external stakeholders, such as suppliers,
local authorities, and NGOs, are essential for extending an organization’s reach and re-
sources during disasters. This interconnectivity is a common feature in organizational
logistics, ensuring that no entity operates in isolation. This clarification confirms why
these centers can be considered to be relief centers.
• Medical Centers: Health and safety is paramount during disaster management. Incor-
porating medical centers ensures that immediate medical needs are met, a necessity
for all organizations involved in humanitarian efforts.
By including these components, the proposed model ensures a comprehensive and
versatile framework that can be adapted to various types of organizations and their specific
needs. This generalizability is intended to provide a robust foundation that can be tailored
to the unique contexts of different organizations, enhancing their capacity to manage
disasters effectively.
In the preceding sections, we conducted several numerical examples and extensive
sensitivity analysis to enhance our comprehension of the proposed model’s behavior. This
endeavor yielded significant economic and managerial implications, which are summarized
as follows:
(1) According to Figures 9–12, managers in the organization are advised that if they
encounter budget limitations, they should ensure that the budget before a disaster is
Systems 2024, 12, 215 25 of 29

maintained at an acceptable level to sufficiently open PRCs. They can then increase
the budget after the disaster occurs. The proposed model can determine the acceptable
level for both budgets.
(2) Based on Figures 13 and 14, the organization mangers should use the advanced AAs
and EMVs to transport the injuries from affected areas to medical centers as soon
as possible. Additionally, effective maintenance of AAs and EMVs is crucial, as it
can significantly impact the health and reliability of these transportation systems.
Figure 13 clearly shows how the speed of the vehicle can affect injured transfer time.
Furthermore, there are various types of AAs available, and the organization can opt
for types with shorter setup times and higher speeds.
(3) Based on Figure 14, total cost can be reduced when the speed increases. Hence, the
organization should consider this saving and invest in renewing the transportation
system. It should be noted the value of this saving can be obtained through the
proposed model.
(4) Considering that weight and wind significantly affect helicopter performance, and
given the importance of helicopter speed in the proposed model, organizations should
assess wind speeds before employing AAs to transfer injured individuals and rescue
groups. Based on this assessment, the allocation of AAs can be determined. Also, the
weight of AAs should be considered before the rescue operation after disaster.
(5) Based on Figures 15 and 16, the decision-makers in organizations should increase
EMVs when the number of AAs are limited, which resulted in decreasing number of
untreated injured and total cost. In this case, they can allocate more AAs to transport
rescue group of medical centers to affected areas.
(6) When more AAs are allocated to transport rescue groups, assuming that EMVs in-
crease, managers should implement transportation planning and routing solutions.
This involves selecting the best and least congested roads for EMVs to enhance their
expected speed and reduce the time that injured individuals spend in EMVs.
(7) Based on Figure 17, the organizations can considerably reduce the total time that
injured individuals spend on roads to arrive medical centers, total cost, and the total
number of untreated injured individuals decreases by increasing the number of AAs.
(8) By comparing Figures 15 and 17, it is concluded that if the number of AAs remain
unchanged, the increasing number of EMVs leads to an increase in the total time that
injured individuals spend on roads to arrive medical centers. Therefore, organiza-
tion managers should consistently allocate funds for acquiring AAs to expedite the
transportation of injured individuals to medical centers. Increasing just EMVs is not a
good strategy to transport patients to medical centers.
(9) Organization managers are advised to increase EMVs or allocate more budget to them
when the number of AAs is at a sufficient level, which can be determined by the
proposed model. In other words, as long as the number of AAs is not enough for
transferring rescue groups to affected areas, an increase in EMVs leads to an increase
in the time that injured individuals spend in roads to arrive medical centers.
(10) According to the results, it is suggested that the organization transfer some staff from
operations centers/field offices and external stakeholders to affected areas for the
distribution of RIs and helping in rescue operations. The proposed model can be used
to obtain the optimal number of transferred staff.

8. Conclusions and Future Directions

This study introduces an organization-centered humanitarian logistics and supply
chain network for both pre- and post-disaster planning. For the first time, the proposed
model considered the internal components of the organization, and was developed based
on the main components of the organization, i.e., regional branches, major transportation
hubs, operational units, operations centers/field offices, external stakeholders, and medical
centers as the affected areas. Also, the affected areas, PRCs, TRCs, operations centers/field
offices, external stakeholders, and medical centers are the main centers that were considered
Systems 2024, 12, 215 26 of 29

in the model. Moreover, the operations centers/field offices and external stakeholders
were considered to be the relief parts. This paper develops a MINLP model to minimize
the total costs associated with pre- and post-disasters. Some new cost factors, such as
delayed total time of rescue and distribution of RIs operations, and waiting cost of injured
persons in EMVs and AAs, have been taken into account. The proposed model was
solved using GAMS software in small and medium scales. Because of the complexity of
the model, the GOA was proposed as a metaheuristic method for large size problems.
The results showed that the GOA has an acceptable performance for solving large-scale
problems. Extensive numerical examples and selectivity analysis were conducted to show
the applicability and behavior of the model. Some important managerial insights for
the manager in organizations were presented. Furthermore, the results indicate that the
organization should use advanced AAs and EMVs for transferring injuries, and implement
an effective maintenance on these transportation systems. Moreover, simply increasing
EMVs is not an appropriate policy for transporting patients to medical centers. Managers
should prioritize investing more in providing AAs.
Although this model has been presented for the first time in an organization, and tried
to consider all aspects, there are some limitations to this study. Several parameters, like the
number of staff required to help with the rescue operations and in distributing RIs, the time
span for the rescue operation or distributing RIs being delayed due to the absence of staff,
the number of available staff in operations centers/field offices and external stakeholders,
etc., were assumed to be deterministic, while considering them to be uncertain parameters
may be more realistic and is recommended in future studies. It has been assumed that
injuries are just transported to medical centers, while in some conditions where there
are a lack of internal medical centers, they can be transported to external clinics and
medical centers, which can be an interesting topic for future directions. Although this paper
considered two types of injuries in the affected area, considering damage severity could be
another research topic in the future. Finally, considering interactions among organizations
to provide a quick response for affected areas is recommended for future research.

Author Contributions: Conceptualization, A.A.; Methodology, A.A.; Software, A.A.; Validation,

A.A.; Formal analysis, A.A.; Investigation, A.A.; Resources, A.A. and N.M.M.; Writing—original
draft, A.A.; Writing—review and editing, S.S., N.M.M., E.H., F.J. and R.Y.; Visualization, A.A., S.S.
and N.M.M.; Supervision, S.S., N.M.M., E.H. and F.J.; Project administration, S.S., N.M.M., E.H., F.J.
and R.Y.; Funding acquisition, R.Y. All authors have read and agreed to the published version of the
manuscript.
Funding: This research received no external funding.
Data Availability Statement: Data are contained within the article.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Wang, J.; Ng, Y.Y.E. Post-earthquake housing recovery with traditional construction: A preliminary review. Prog. Disaster Sci.
2023, 18, 100283. [CrossRef]
2. Kotani, H.; Okai, H.; Tamura, M. Activities and roles of mosques in Japan after the recent major earthquakes: A comprehensive
study. Prog. Disaster Sci. 2023, 20, 100297. [CrossRef]
3. Kodag, S.; Mani, S.K.; Balamurugan, G.; Bera, S. Earthquake and flood resilience through spatial Planning in the complex urban
system. Prog. Disaster Sci. 2022, 14, 100219. [CrossRef]
4. Mukherjee, M.; Abhinay, K.; Rahman, M.M.; Yangdhen, S.; Sen, S.; Adhikari, B.R.; Nianthi, R.; Sachdev, S.; Shaw, R. Extent and
evaluation of critical infrastructure, the status of resilience and its future dimensions in South Asia. Prog. Disaster Sci. 2023, 17,
100275. [CrossRef]
5. Jihad, A.; Muksin, U.; Syamsidik; Ramli, M.; Banyunegoro, V.H.; Simanjuntak, A.V.H.; Rusdin, A.A. Tsunami evacuation sites in
the northern Sumatra (Indonesia) determined based on the updated tsunami numerical simulations. Prog. Disaster Sci. 2023, 18,
100286. [CrossRef]
6. Marome, W.; Rodkul, P.; Mitra, B.K.; Dasgupta, R.; Kataoka, Y. Towards a more sustainable and resilient future: Applying the
Regional Circulating and Ecological Sphere (R-CES) concept to Udon Thani City Region, Thailand. Prog. Disaster Sci. 2022, 14,
100225. [CrossRef]
Systems 2024, 12, 215 27 of 29

7. Gray, J.; Lloyd, S.; Healey, S.; Opdyke, A. Urban and rural patterns of typhoon mortality in the Philippines. Prog. Disaster Sci.
2022, 14, 100234. [CrossRef]
8. Ginige, K.; Mendis, K.; Thayaparan, M. An assessment of structural measures for risk reduction of hydrometeorological disasters
in Sri Lanka. Prog. Disaster Sci. 2022, 14, 100232. [CrossRef]
9. Hoeppe, P. Trends in weather related disasters–Consequences for insurers and society. Weather Clim. Extrem. 2016, 11, 70–79.
[CrossRef]
10. Chaudhary, M.T.; Piracha, A. Natural disasters—Origins, impacts, management. Encyclopedia 2021, 1, 1101–1131. [CrossRef]
11. Hwang, K.-E.; Kim, I. Post-COVID-19 modular building review on problem-seeking framework: Function, form, economy, and
time. J. Comput. Des. Eng. 2022, 9, 1369–1387. [CrossRef]
12. Farazmehr, S.; Wu, Y. Locating and deploying essential goods and equipment in disasters using AI-enabled approaches: A
systematic literature review. Prog. Disaster Sci. 2023, 19, 100292. [CrossRef]
13. Bodin, Ö.; Nohrstedt, D.; Orach, K. A diagnostic for evaluating collaborative responses to compound emergencies. Prog. Disaster
Sci. 2022, 16, 100251. [CrossRef]
14. Daneshvar, M.; Jena, S.D.; Rei, W. A two-stage stochastic post-disaster humanitarian supply chain network design problem.
Comput. Ind. Eng. 2023, 183, 109459. [CrossRef]
15. Anjomshoae, A.; Banomyong, R.; Mohammed, F.; Kunz, N. A systematic review of humanitarian supply chains performance
measurement literature from 2007 to 2021. Int. J. Disaster Risk Reduct. 2022, 72, 102852. [CrossRef]
16. McLoughlin, D. A framework for integrated emergency management. Public Adm. Rev. 1985, 45, 165–172. [CrossRef]
17. Kovács, G.; Spens, K.M. Humanitarian logistics in disaster relief operations. Int. J. Phys. Distrib. Logist. Manag. 2007, 37, 99–114.
[CrossRef]
18. Balcik, B.; Bozkir, C.D.C.; Kundakcioglu, O.E. A literature review on inventory management in humanitarian supply chains. Surv.
Oper. Res. Manag. Sci. 2016, 21, 101–116. [CrossRef]
19. Yi, W.; Özdamar, L. A dynamic logistics coordination model for evacuation and support in disaster response activities. Eur. J.
Oper. Res. 2007, 179, 1177–1193. [CrossRef]
20. Berkoune, D.; Renaud, J.; Rekik, M.; Ruiz, A. Transportation in disaster response operations. Socio-Econ. Plan. Sci. 2012, 46, 23–32.
[CrossRef]
21. Rawls, C.G.; Turnquist, M.A. Pre-positioning of emergency supplies for disaster response. Transp. Res. Part B Methodol. 2010, 44,
521–534. [CrossRef]
22. Döyen, A.; Aras, N.; Barbarosoğlu, G. A two-echelon stochastic facility location model for humanitarian relief logistics. Optim.
Lett. 2012, 6, 1123–1145. [CrossRef]
23. Najafi, M.; Eshghi, K.; de Leeuw, S. A dynamic dispatching and routing model to plan/re-plan logistics activities in response to
an earthquake. OR Spectr. 2014, 36, 323–356. [CrossRef]
24. Camacho-Vallejo, J.-F.; González-Rodríguez, E.; Almaguer, F.-J.; González-Ramírez, R.G. A bi-level optimization model for aid
distribution after the occurrence of a disaster. J. Clean. Prod. 2015, 105, 134–145. [CrossRef]
25. Rezaei-Malek, M.; Tavakkoli-Moghaddam, R.; Zahiri, B.; Bozorgi-Amiri, A. An interactive approach for designing a robust
disaster relief logistics network with perishable commodities. Comput. Ind. Eng. 2016, 94, 201–215. [CrossRef]
26. Tofighi, S.; Torabi, S.A.; Mansouri, S.A. Humanitarian logistics network design under mixed uncertainty. Eur. J. Oper. Res. 2016,
250, 239–250. [CrossRef]
27. Hu, S.-L.; Han, C.-F.; Meng, L.-P. Stochastic optimization for joint decision making of inventory and procurement in humanitarian
relief. Comput. Ind. Eng. 2017, 111, 39–49. [CrossRef]
28. Noham, R.; Tzur, M. Designing humanitarian supply chains by incorporating actual post-disaster decisions. Eur. J. Oper. Res.
2018, 265, 1064–1077. [CrossRef]
29. Vahdani, B.; Veysmoradi, D.; Noori, F.; Mansour, F. Two-stage multi-objective location-routing-inventory model for humanitarian
logistics network design under uncertainty. Int. J. Disaster Risk Reduct. 2018, 27, 290–306. [CrossRef]
30. Liu, Y.; Lei, H.; Wu, Z.; Zhang, D. A robust model predictive control approach for post-disaster relief distribution. Comput. Ind.
Eng. 2019, 135, 1253–1270. [CrossRef]
31. Momeni, B.; Aghsami, A.; Rabbani, M. Designing humanitarian relief supply chains by considering the reliability of route, repair
groups and monitoring route. Adv. Ind. Eng. 2019, 53, 93–126.
32. Sakiani, R.; Seifi, A.; Khorshiddoust, R.R. Inventory routing and dynamic redistribution of relief goods in post-disaster operations.
Comput. Ind. Eng. 2020, 140, 106219. [CrossRef]
33. Aghajani, M.; Torabi, S.A.; Heydari, J. A novel option contract integrated with supplier selection and inventory prepositioning for
humanitarian relief supply chains. Socio-Econ. Plan. Sci. 2020, 71, 100780. [CrossRef]
34. Heydari, H.; Aghsami, A.; Rabani, M. A mathematical model to optimize debris clearance problem in the disaster response Phase:
A case study. J. Ind. Syst. Eng. 2021, 14, 1–34.
35. Abazari, S.R.; Aghsami, A.; Rabbani, M. Prepositioning and distributing relief items in humanitarian logistics with uncertain
parameters. Socio-Econ. Plan. Sci. 2021, 74, 100933. [CrossRef]
36. Abazari, S.R.; Jolai, F.; Aghsami, A. Designing a humanitarian relief network considering governmental and non-governmental
operations under uncertainty. Int. J. Syst. Assur. Eng. Manag. 2021, 13, 1430–1452. [CrossRef]
Systems 2024, 12, 215 28 of 29

37. Hajipour, V.; Niaki, S.T.A.; Akhgar, M.; Ansari, M. The healthcare supply chain network design with traceability: A novel
algorithm. Comput. Ind. Eng. 2021, 161, 107661. [CrossRef]
38. Masoumi, M.; Aghsami, A.; Alipour-Vaezi, M.; Jolai, F.; Esmailifar, B. An M/M/C/K queueing system in an inventory routing
problem considering congestion and response time for post-disaster humanitarian relief: A case study. J. Humanit. Logist. Supply
Chain Manag. 2022, 12, 182–219. [CrossRef]
39. Bakhshi, A.; Aghsami, A.; Rabbani, M. A scenario-based collaborative problem for a relief supply chain during post-disaster
under uncertain parameters: A real case study in Dorud. J. Model. Manag. 2022, 18, 906–941. [CrossRef]
40. Ehsani, B.; Karimi, H.; Bakhshi, A.; Aghsami, A.; Rabbani, M. Designing humanitarian logistics network for managing epidemic
outbreaks in disasters using Internet-of-Things. A case study: An earthquake in Salas-e-Babajani city. Comput. Ind. Eng. 2023, 175,
108821. [CrossRef]
41. Sheikholeslami, M.; Zarrinpoor, N. Designing an integrated humanitarian logistics network for the preparedness and response
phases under uncertainty. Socio-Econ. Plan. Sci. 2023, 86, 101496. [CrossRef]
42. Lohrasbpoor, M.; Arshadi Khamseh, A.; Nemati-Lafmejani, R.; Naderi, B. Four Echelons Humanitarian Network Design
Considering Capacitated/lateral Transshipment with a Destruction Radius and ABO Compatibility: Tehran Earthquake. Int. J.
Supply Oper. Manag. 2023, 10, 1–22.
43. Khalili-Fard, A.; Hashemi, M.; Bakhshi, A.; Yazdani, M.; Jolai, F.; Aghsami, A. Integrated relief pre-positioning and procurement
planning considering non-governmental organizations support and perishable relief items in a humanitarian supply chain
network. Omega 2024, 127, 103111. [CrossRef]
44. Altay, N.; Heaslip, G.; Kovács, G.; Spens, K.; Tatham, P.; Vaillancourt, A. Innovation in humanitarian logistics and supply chain
management: A systematic review. Ann. Oper. Res. 2024, 335, 965–987. [CrossRef] [PubMed]
45. Tarei, P.K.; Gumte, K.M.; Patnaik, J.; Oktari, R.S. Analysing barriers to humanitarian logistics for distributing relief aid in pre-and
post-disaster situations. Int. J. Disaster Risk Reduct. 2024, 104, 104388. [CrossRef]
46. Beiki Ashkezari, A.; Zokaee, M.; Rabbani, E.; Rabbani, M.; Aghsami, A. A scenario-based game theory integrating with a
location-allocation-routing problem in a pre-and post-disaster humanitarian logistics network under uncertainty. J. Model. Manag.
2024. [CrossRef]
47. Mousavi, S.; Sajadi, S.M.; AlemTabriz, A.; Najafi, S.E. Hybrid mathematical and simulation model for designing a hierarchical
network of temporary medical centers in a disaster. J. Simul. 2024, 18, 119–135. [CrossRef]
48. Liu, B.; Zhou, Y.; Luo, Q.; Huang, H. Quantum-inspired African vultures optimization algorithm with elite mutation strategy for
production scheduling problems. J. Comput. Des. Eng. 2023, 10, 1767–1789. [CrossRef]
49. Ma, B.; Hu, Y.; Lu, P.; Liu, Y. Running city game optimizer: A game-based metaheuristic optimization algorithm for global
optimization. J. Comput. Des. Eng. 2023, 10, 65–107. [CrossRef]
50. Chen, S.; Zheng, J. Sand cat arithmetic optimization algorithm for global optimization engineering design problems. J. Comput.
Des. Eng. 2023, 10, 2122–2146. [CrossRef]
51. Goodarzimehr, V.; Talatahari, S.; Shojaee, S.; Hamzehei-Javaran, S.; Sareh, P. Structural design with dynamic constraints using
weighted chaos game optimization. J. Comput. Des. Eng. 2022, 9, 2271–2296. [CrossRef]
52. Xing, J.; Zhao, Q.; Chen, H.; Zhang, Y.; Zhou, F.; Zhao, H. Utilizing bee foraging behavior in mutational salp swarm for feature
selection: A study on return-intentions of overseas Chinese after COVID-19. J. Comput. Des. Eng. 2023, 10, 2094–2121. [CrossRef]
53. Jia, H.; Wen, Q.; Wu, D.; Wang, Z.; Wang, Y.; Wen, C.; Abualigah, L. Modified beluga whale optimization with multi-strategies for
solving engineering problems. J. Comput. Des. Eng. 2023, 10, 2065–2093. [CrossRef]
54. Eslami, N.; Yazdani, S.; Mirzaei, M.; Hadavandi, E. A modified binary version of aphid-ant mutualism for feature selection: A
COVID-19 case study. J. Comput. Des. Eng. 2023, 10, 549–577. [CrossRef]
55. Wang, S.; Hussien, A.G.; Kumar, S.; AlShourbaji, I.; Hashim, F.A. A modified smell agent optimization for global optimization
and industrial engineering design problems. J. Comput. Des. Eng. 2023, 10, 2147–2176. [CrossRef]
56. Premkumar, M.; Jangir, P.; Sowmya, R.; Alhelou, H.H.; Mirjalili, S.; Kumar, B.S. Multi-objective equilibrium optimizer: Framework
and development for solving multi-objective optimization problems. J. Comput. Des. Eng. 2022, 9, 24–50. [CrossRef]
57. Zheng, R.; Hussien, A.G.; Qaddoura, R.; Jia, H.; Abualigah, L.; Wang, S.; Saber, A. A multi-strategy enhanced African vultures
optimization algorithm for global optimization problems. J. Comput. Des. Eng. 2023, 10, 329–356. [CrossRef]
58. Jia, H.; Li, Y.; Wu, D.; Rao, H.; Wen, C.; Abualigah, L. Multi-strategy Remora Optimization Algorithm for solving multi-extremum
problems. J. Comput. Des. Eng. 2023, 10, 1315–1349. [CrossRef]
59. Ylldlz, B.S.; Mehta, P.; Panagant, N.; Mirjalili, S.; Yildiz, A.R. A novel chaotic Runge Kutta optimization algorithm for solving
constrained engineering problems. J. Comput. Des. Eng. 2022, 9, 2452–2465. [CrossRef]
60. Gorji, S.A. Challenges and opportunities in green hydrogen supply chain through metaheuristic optimization. J. Comput. Des.
Eng. 2023, 10, 1143–1157. [CrossRef]
61. Mousapour Mamoudan, M.; Ostadi, A.; Pourkhodabakhsh, N.; Fathollahi-Fard, A.M.; Soleimani, F. Hybrid neural network-based
metaheuristics for prediction of financial markets: A case study on global gold market. J. Comput. Des. Eng. 2023, 10, 1110–1125.
[CrossRef]
62. Jia, H.; You, F.; Wu, D.; Rao, H.; Wu, H.; Abualigah, L. Improved snow ablation optimizer with heat transfer and condensation
strategy for global optimization problem. J. Comput. Des. Eng. 2023, 10, 2177–2199. [CrossRef]
Systems 2024, 12, 215 29 of 29

63. Qiao, S.; Yu, H.; Heidari, A.A.; El-Saleh, A.A.; Cai, Z.; Xu, X.; Mafarja, M.; Chen, H. Individual disturbance and neighborhood
mutation search enhanced whale optimization: Performance design for engineering problems. J. Comput. Des. Eng. 2022, 9,
1817–1851. [CrossRef]
64. Chen, Y.; Lu, C.; Yan, J.; Feng, J.; Sareh, P. Intelligent computational design of scalene-faceted flat-foldable tessellations. J. Comput.
Des. Eng. 2022, 9, 1765–1774. [CrossRef]
65. Sogandi, F.; Shiri, M. Metaheuristic algorithms for a sustainable saffron supply chain network considering government policies
and product quality under uncertainty. J. Comput. Des. Eng. 2023, 10, 1892–1929. [CrossRef]
66. Cho, Y.I.; Nam, S.H.; Cho, K.Y.; Yoon, H.C.; Woo, J.H. Minimize makespan of permutation flowshop using pointer network. J.
Comput. Des. Eng. 2022, 9, 51–67. [CrossRef]
67. Saremi, S.; Mirjalili, S.; Lewis, A. Grasshopper optimisation algorithm: Theory and application. Adv. Eng. Softw. 2017, 105, 30–47.
[CrossRef]
68. Gouran-Orimi, S.; Ghasemi-Marzbali, A. Load Frequency Control of multi-area multi-source system with nonlinear structures
using modified Grasshopper Optimization Algorithm. Appl. Soft Comput. 2023, 137, 110135. [CrossRef]
69. Alrashidi, M.; Alrashidi, M.; Rahman, S. Global solar radiation prediction: Application of novel hybrid data-driven model. Appl.
Soft Comput. 2021, 112, 107768. [CrossRef]
70. Ingle, K.K.; Jatoth, R.K. Non-linear Channel Equalization using Modified Grasshopper Optimization Algorithm. Appl. Soft
Comput. 2024, 153, 110091. [CrossRef]
71. Hirbod, F.; Eshghali, M.; Sheikhasadi, M.; Jolai, F.; Aghsami, A. A state-dependent M/M/1 queueing location-allocation model
for vaccine distribution using metaheuristic algorithms. J. Comput. Des. Eng. 2023, 10, 1507–1530. [CrossRef]
72. Abbaspour, S.; Aghsami, A.; Jolai, F.; Yazdani, M. An integrated queueing-inventory-routing problem in a green dual-channel
supply chain considering pricing and delivery period: A case study of construction material supplier. J. Comput. Des. Eng. 2022, 9,
1917–1951. [CrossRef]
73. Luo, J.; Chen, H.; Xu, Y.; Huang, H.; Zhao, X. An improved grasshopper optimization algorithm with application to financial
stress prediction. Appl. Math. Model. 2018, 64, 654–668. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.