Multicriteria optimization approach to deploy humanitarian logistic operations integrally during floods

• Author: Julián Molina,Begoña Vitoriano,Christopher Mejia‐Argueta,Juan Gaytán,Rafael Caballero
• Published date: 01 May 2018
• DOI: http://doi.org/10.1111/itor.12508
• Date: 01 May 2018

Intl. Trans. in Op. Res. 25 (2018) 1053–1079

DOI: 10.1111/itor.12508

INTERNATIONAL

TRANSACTIONS

IN OPERATIONAL

RESEARCH

Multicriteria optimization approach to deploy humanitarian

logistic operations integrally during floods

Christopher Mejia-Arguetaa, Juan Gayt´

anb, Rafael Caballeroc,Juli

´

an Molinacand

Bego˜

na Vitorianod

aDepartment of Industrial Engineering and Innovation Sciences, Technische UniversiteitEindhoven, 3169, Den Dolech 2,

Eindhoven, Nord Brabant5612AZ, the Netherlands

bSchool of Engineering, Universidad Autonoma del Estado de Mexico,225104, Toluca, Estado de Mexico, Mexico

cApplied Economics (Mathematics), Campus El Ejido, University of M´

alaga, M´

alaga 29071, Spain

dDepartment of Statistics and Operational Research and Institute of Interdisciplinary Mathematics, Comunidad de

Madrid, Universidad Complutense de Madrid, 16734 Madrid, Spain

E-mail: christopher.mejia.argueta@gmail.com [Mejia-Argueta]; jgi@uaemex.mx [Gayt´

an];

r_caballero@uma.es [Caballero]; julian.molina@uma.es [Molina]; bvitoriano@mat.ucm.es [Vitoriano]

Received 13 October 2016; receivedin revised form 15 December 2017; accepted 23 December 2017

Abstract

This paper addresses frequent and foreseeable floods in the short-term preparedness of an imminent event

using a multicriteria optimization model integrated with a geographical information system to simulateflood

levels, determine the best strategies, and update information. The proposed model takes into account the

four main relief operations: location of emergency facilities (i.e., distribution centers, shelters, and meeting

points), prepositioning of humanitarian aid, evacuation, and distribution of humanitarian aid. Threecriteria

are considered in the formulation to minimize: the maximum evacuation flow-time, the maximum distribu-

tion flow-time, and total cost of relief operations. The approximation to the efficient frontier is built using

multiobjective programming through the use of commercial software. The usefulness and robustness of the

model are verified using data from one of the worst Mexican floods considering various flood levels created

from three key elements in humanitarian logistics. The strategies provided by the proposed methodology are

compared with those implemented by the Mexican authorities during the studied disaster.

Keywords:multiple-objective programming; efficient solution; weighted-sum method; ε-constraint; humanitarian opera-

tions

1. Introduction

Natural disasters have impacted the world’s population throughout the history of humanity with

terrible consequences for inhabitants and their environment, as observed in several disasters over

the past decade reported in Aon Benfield (2016). According to the International Database of

C

2018 The Authors.

International Transactionsin Operational Research C

2018 International Federation of OperationalResearch Societies

Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA02148,

USA.

1054 C. Mejia-Argueta et al. / Intl. Trans.in Op. Res. 25 (2018) 1053–1079

Disasters EM-DAT (2016), the number of disasters affecting countries around the world is appar-

ently increasing, as well as the number of people affected by them (approximately 4.8 billion people

were affected by naturaldisasters between 1970 and 2003 compared to around two billion people in

the last decade). However, the numberof victims is decreasing, showing the ability of the community

to protect itself and increase its resilience.

Among natural disasters, the maximum percentage increase is shown by hydrometeorological

disasters, which constituted a share of 82% during the last decade (23% more than in the period

1970–2003) and affected almost one billion people. These disasters are closely related to seasonal

weather events and can be accurately simulated in time, location, and magnitude, allowing for the

growth of more effective plans to address their consequences (D´

ıaz-Delgado and Gayt´

an, 2014).

Disaster management is related to planning, implementing, and controlling effective strategies to

alleviate human sufferingand reduce negative effects of disasters.The so-called disaster management

cycle is divided into four phases (Tomasini and Van Wassenhove, 2009): mitigation, preparedness,

response, and recovery/reconstruction. Despite the knowledge gained during the last two decades,

disaster management remains a major challenge, creating important research opportunities in the

analysis of integrated humanitarian operations and the application of multiple-criteria decision

making (see Ortu˜

no et al., 2013; Leiras et al., 2014; Gutjahr and Nolz, 2016).

Multiobjective optimization is a research field that has grown since the end of the last century

and it is gaining more traction given the opportunities to analyze tradeoffs of multiple criteria

in the same model (Ehrgott, 2005). In general, the related techniques provide a decision maker

the opportunity to identify and evaluate various alternative high-quality approximations to optimal

solutions (nondominated or Paretooptima) in order to support her final decision. This is particularly

useful for decision makers in humanitarian contexts where there are diverse conflicting criteria in

the operations. Furthermore, stakeholders assess the scarce resources or try to meet specific values,

then they need to find the most suitable solution, and multicriteria optimizationprovides them this.

The aim of our approach is to introduce a methodology to make better decisions during the

disaster preparedness phase, when the event is about to occur and becomes an emergency. This

methodology involves two phases: (a) a geographical information system (GIS) that is used to

simulate flood maps and evaluate damage in the available infrastructure (i.e., road network and

potential emergency buildings)and (b) a multiobjective optimization model to determine the number

and location of emergency facilities to be opened and the flow of evacuees and humanitarian aid

through the availablenetwork using multiple vehicles,taking into account several criteria: evacuation

and distribution flow-time, budget usage in various flood cases.

Therefore, the main contribution of this paper with regard to similar studies (e.g., Rodriguez-

Espindola and Gaytan, 2015) is the formulation of a multicriteria optimization model that con-

templates a novel approach in evacuation using a two-tiered strategy via meeting points, considers

infrastructure saturation and availabilityof resources (i.e., vehicles, budget, facilities), and minimize

the worst-case scenario to perform people evacuation and distribution of relief products under

diverse circumstances in the short term after the disaster occurs. The paper is organized as follows.

In Section 2, a brief literature review is presented. In Section 3, the framework of the proposed

methodology is described together with the mathematical formulation of the problem under study.

Once the methodology and the model are described, the results of a real Mexican case study related

to a large flood in 2007 and a set of test instances are discussed in Section 4. Finally, Section 5

presents the study’s conclusions and a number of suggestions for future work.

C

2018 The Authors.

International Transactionsin Operational Research C

2018 International Federation of OperationalResearch Societies