Multi-Criteria Decision Making in Disaster Preparedness and Response

Sibel Salman, Department of Industrial Engineering, Koç University

Abstract

Disaster preparedness and response operations are characterized  by uncertainty, large-scale impact, and the  involvement  of  multiple  decision-makers  having  multiple  criteria.  In  pre-disaster  preparedness,  decisions  involving the location of emergency response facilities, design of a relief distribution system, prepositioning  of  relief  aid  inventory,  and  strengthening  of  infrastructure  networks  need  to  be  taken  under  uncertainty.  In  post-disaster response, decisions involve dispatching and routing of first responders, transporting casualties to  medical emergency units as well as vital first-aid commodities and emergency personnel to disaster-affected  areas,  evacuation  of  people,  and  clearance  of  debris  in  a  dynamic  setting  under  evolving  information.  The  objectives optimized are classified under efficiency, effectiveness, and equity. Equity, which aims for the equal  distribution of benefits or disutilities, is a major concern in response services where resources are scarce, and  timeliness is critical. In the disaster response context, equity attributes such as demand satisfaction rate, arrival  time, deprivation cost, travel distance, accessibility, and service level have received attention in the literature.  As an example, we consider a stochastic shelter location problem arising in the preparedness stage. We take  both  efficiency  and  inequity  objectives  into  account  by  minimizing  a  linear  combination  of  (i)  the  mean  distance between opened shelter locations and the locations of the individuals assigned to them, and (ii) Gini’s  Mean Absolute Difference of these distances. Furthermore, a chance constraint is defined on the total cost of  opening the shelters and their capacity expansion. We develop a stochastic programming model with scenarios  for uncertain demand and disruptions in the transportation network, and a Genetic Algorithm (GA) that utilizes  a mixed-integer programming subproblem. The GA yields small optimality gaps in a short time for benchmark  instances. We run the GA also on Istanbul data to drive insights to guide decision-makers.

Short Bio

Sibel  Salman  is  a  Professor  at  the  Industrial  Engineering  Department  of  the  College  of  Engineering  at  Koç  University in Istanbul, Turkey.  Prior to joining Koç University, she held a faculty position at the Krannert School  of  Management,  Purdue  University,  USA.  She  got  her  Ph.D.  in  Operations  Research  from  Carnegie  Mellon  University, USA, and her M.Sc. and B.Sc. degrees from Bilkent University, Turkey. She has published in the areas  of disaster and health care logistics, supply chain management, facility location, vehicle routing, network design,  production scheduling, approximation algorithms, and combinatorial optimization. She is a department editor for  OR Spectrum and an associate editor for Sustainable Analytics and Modeling journals and has been/is in the editorial  board of Computers and Operations Research and Production and Operations Management journals. She is on the  board of EURO working group on Humanitarian Operations (HOpe) and a member of the EURO WISDOM Forum.  Her  current  research  interests  are  in  humanitarian  logistics,  disaster  management,  healthcare  logistics,  facility  location, vehicle routing.

Venue

Friday, November 5, 2021, 4.00 pm - Zoom Meeting