Available online at www.sciencedirect.com ScienceDirect Procedia Engineering 78 (2014 ) 59 63 Humanitarian Technology: Science, Systems and Global Impact 2014, HumTech2014 Multi-criteria decision analysis approach to site suitability of U.S. Department of Defense humanitarian assistance projects Richard W. Curran a *, Matthew E. Bates b, Heather M. Bell c a US Army Corps of Engineers, Engineer Research and Development Center, Topographic Engineering Center, Alexandria, VA, USA b US Army Corps of Engineers, Engineer Research and Development Center, Environmental Lab, Concord, MA, USA c Pacific Disaster Center, Kihei, HI, USA Abstract This paper seeks to provide a framework for determining best locations for US Department of Defense (DOD) Humanitarian Assistance (HA) projects as well as proposing a method to selecting and optimal portfolio of DOD projects for an area of operation through data-driven analyses and an understanding of priorities and tradeoffs. To identify optimal locations for HA projects a Multi-Criteria Decision Analysis (MCDA) approach to site suitability analysis is presented along with a case-study using data from El Salvador to demonstrate the utility of the approach. A Portfolio Decision Analysis (PDA) is proposed as a method to generate a portfolio with the most optimal combination of HA projects to lower resource investment risk and maximize desired impact. These two methods are intended as decision aids for HA managers and project nominators in the DOD HA project approval process. Public Release, Distribution Unlimited Published 2014 The by Authors. Elsevier Published Ltd. This is by an Elsevier open access Ltd. article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of the Organizing Committee of HumTech2014. Selection and peer-review under responsibility of the Organizing Committee of HumTech2014 Keywords: MCDA, Site Suitability, Humanitarian Assistance 1. Introduction The U.S. Department of Defense (DoD) utilizes humanitarian assistance (HA) as a way to create positive relationships with partner nations. DoD guidance [1] has led to a focus on non-kinetic operations to help create more stable social environments with countries that have cooperative relationships with the US DOD and other US * Corresponding author. Tel.: +703.428.3625; fax: +703.428.3732 E-mail address: Richard.W.Curran@usace.army.mil 1877-7058 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of the Organizing Committee of HumTech2014 doi:10.1016/j.proeng.2014.07.039
60 Richard W. Curran et al. / Procedia Engineering 78 ( 2014 ) 59 63 Government entities. Civilian-Military interactions are often the most effective method used to sustain stable environments and HA operations are the enabler of such stability. It is important for these Civil-Military Operations (CMO) to support partner nation governments and build civilian capacity to respond to emergency situations and build resilience to natural disasters. As a result, requirements exist for DOD-sponsored efforts to have a high probability of achieving desired outcomes and meeting performance expectations outlined by both the US and the host nation. Additionally, increased fiscal scrutiny has pushed the DOD to better ensure proper and effective use of tax-payer funds. The HA resource-allocation decision process that DOD HA managers go through is inherently a multi-criteria decision analysis (MCDA) problem. An HA manager weighs the importance of various sources of information, location specific data, DOD strategic guidance, host nation priorities, other US government agencies HA activities and their own personal expertise to determine which new projects should be approved. An MCDA site suitability decision tool that accounts for these types of information, data, priorities, and user expertise is needed to provide documentation, transparency, and analytics to the project nomination process and help validate decisions. This paper applies the proposed decision-analytic framework to DOD HA activities funded by Overseas Humanitarian, Disaster Assistance, and Civic Aid (OHDACA), which seeks to promote social stability and resilience to disaster incidents within partner countries [2]. 2. Multi-Criteria Decision Analysis (MCDA) MCDA is a long standing field that has developed over several decades to provide structure to the decision space and a way to prioritize and evaluate numerous, sometimes conflicting, decision criteria to evaluate and rank alternatives [3]. A strength of MCDA is its ability to incorporate both qualitative and quantitative criteria in the decision space [4]. MCDA is agnostic to the type of decision space, so application to geospatial decisions is logical and seamless. In fact, geographical information systems (GIS) based MCDA analyses have been implemented since the early 1990 s [3,5,6]. A structure of criteria or objective statements and user-determined weights identifying the relative importance of the criteria are defined by the user in the MCDA. The MCDA decision logic analyzes the alternatives with respect to the criteria and their respective weights. In this paper, two methods are used; Multi- Attribute Value Theory (MAVT) and Preference Ranking Organization Method for Enriching Evaulations II (PROMETHEE II). MAVT is a more quantitative method that uses mathematical value functions to normalize alternative scores before aggregating across criteria to determine a total suitability score for each alternative [7]. PROMETHEE II is a more qualitative method which utilizes preference functions and minimum thresholds for defining outranking relationships between alternatives in a preferred order [8]. Using these methods the MCDA site suitability approach can be used for project evaluations in the nomination process for new OHDACA HA projects. 2.1. Criteria Structure The Pacific Disaster Center performs Risk and Vulnerability Assessments (RVA) for countries to identify where they are most vulnerable to hazards. The RVA is backed by extensive social science and humanitarian assistance domain research, which enables it to characterize areas of operation and help determine where resources are needed to aid at-risk populations and build coping capacity to natural and manmade hazards and emergencies. The RVA framework can be leveraged to provide an MCDA site suitability criteria structure. The criteria structure is organized under four main criteria themes: Multi-Hazard Exposure, Resilience Factors, Investment Sustainability, and Strategic Investment Filter. The four criteria themes and their respective sub-criteria are broken down in Fig. 1. The subcriteria under the Multi-Hazard Exposure and Resilience Factors criteria themes address the question of "where is assistance is most needed?" The Investment Sustainability sub-criteria seek to identify where assistance will be most effective and the Strategic Investment Filter sub-criteria are designed to account for HA project efficiency. The Strategic Investment Filter requires user-specific guidance to be implemented and is still in development: the other three criteria themes populated and included in the use case scenario.
Richard W. Curran et al. / Procedia Engineering 78 ( 2014 ) 59 63 61 Fig. 1. Criteria structure of the HA project MCDA site suitability analysis. The structure is broken out into four criteria themes and sub-criteria for each respective theme. 3. Portfolio Decision Analysis (PDA) The MCDA site suitability methods proposed in this effort apply to individual project evaluation completed by HA project nominators and result in a ranked list of proposed HA locations suitable for each project. Once this nomination process is completed, HA managers must select a portfolio of projects that will best accomplish their desired objectives. A PDA method such as the one introduced by Keisler and Linkov [9] can further be applied to determine an optimal combination of individual projects/locations based on higher-order system constraints (such as diversification, risk and consequences represented by indicators of effectiveness, and total project cost) and project/location correlations. This effort will evaluate various PDA approaches, as well as hard constraints (e.g. total cost) and concepts of risk and consequence as related to HA interventions. The end goal of the applied PDA is to produce recommendations for an optimized portfolio that reduces risk of identified negative consequences through an evaluation of interaction of individual project risks. Additional focus will be placed on characterizing spatial dependences of HA projects in a portfolio. This will make assertions about the appropriate measures of spatial density and complementarity of HA projects by type (e.g. health, infrastructure, disaster mitigation and preparedness, education), which can inform constraints applied during the PDA. 4. Humanitarian Assistance MCDA Site Suitability Use-Case Scenario As a proof-of concept regarding the applicability of MCDA to HA site suitability decisions, we apply two MCDA methods to assess location suitability for a new health initiative in El Salvador. Location suitability data are acquired at the provincial level, with geospatial coverage of all 14 departments of El Salvador. These data used are mainly derived from publically available sources including El Salvador government agencies, U.S. government agencies, international organizations, academic sources and OpenStreetMap. Temporal coverage of the data span
62 Richard W. Curran et al. / Procedia Engineering 78 ( 2014 ) 59 63 from 2007 to 2012, with the exception of the natural hazard data, which ranges from 1900 to 2012. The criteria decision tree for the use-case is shown in Fig. 2. User specified weights identify tradeoffs in priority between these criteria. Fig. 2. Criteria decision tree developed for the El Salvador use-case scenario. The yellow boxes show the four levels of criteria that represent the HA project decision space and the blue boxes identify the 14 alternative locations that are being assessed, which are the departments of El Salvador. Each MCDA method produces a ranked list of alternatives (El Salvador s 14 departments) showing how each department scores relative to the others that can be displayed on a map or in a graph as shown in Figure 3. A comparison of the ranked list of alternatives produced by MAVT and PROMETHEE II is performed to assess result sensitivity to MCDA method choice and or change in criteria weighting. This type of comparison further validates and gives confidence to the MCDA site suitability results.
Richard W. Curran et al. / Procedia Engineering 78 ( 2014 ) 59 63 63 a b Fig. 3. (a) Example of MCDA site suitability results displayed on a map of El Salvador showing the rank of each department by numerical text and by color gradient from green (high) to red (low). (b) The same results in a graph format that shows the rank and score for each department. 5. Summary With the advancement of current technology and analysis techniques and the increased demand for justification of effective use of government resources, opportunities exist to improve the HA resource allocation process. This project introduced in this paper seeks to supply HA managers with a decision aiding tool to inform decisions on where new HA projects should be located as well as identifying which combination of new proposed projects will generate a portfolio that reduces HA resource investment risk and increases the likelihood of achieving desired outputs. A use-case scenario is used to showcase the MCDA site suitability capability developed in an on-going research effort. The incorporation of a PDA framework into the HA project proposal process is also proposed. References [1] Department of Defense, Department of Defense Instruction (Number 3000.05), Retrieved from http://www.dtic.mil/whs/directives/corres/pdf/300005p.pdf, 2009. [2] Defense Security Cooperation Agency, Security Assistance Management Manual (5105.38-M) Chapter 12, Retrieved from http://www.samm.dsca.mil/chapter/chapter-12, 2012. [3] J. Malczewski, GIS-based Multicriteria Decision Analysis: a Survey of the Literature, Int. J. of Geographical Information Science, Vol. 20, No. 7 (2006) 703 726. [4] R. Greene, R. Devillers, J.E. Luther, B.G. Eddy, GIS-based multiple-criteria decision analysis, Geography Compass 5/6 (2011) 412-432. [5] S.J. Carver, Integrating multi-criteria evaluation with geographical information systems, Int. J. Geographical Information Systems, Vol. 5, No. 3 (1991) 321-339. [6] J.M.C. Pereira, L. Duckstein, A multiple criteria decision-making approach to GIS-based land suitability evaluation, Int. J. Geographical Information Systems, Vol. 7, No.53 (1993) 407-424. [7] R. Keeney, H. Raiffa, Decisions with Multiple Objectives: Preferences and Value Trade-Offs. Wiley: New York, 1976. [8] M. Behzadian, R.B. Kazemzadeh, A. Albadvi, M. Aghdasi, PROMETHEE: A comprehensive literature review on methodologies and applications, European J. of Operational Research, 200 (2010) 198-215. [9] J.M. Keisler, I. I. Linkov, Managing A Portfolio of Risks, Wiley Encyclopedia of Operations Research and Management Science, 2010.