Issue 7.2, August 2003
Home

Multicriterial Analysis Application in Mine Action

<-- Back

Click here to view Figure 3

Figure 3 shows a part of the layout displaying density of the identified mines in the polygons that are defined according to their presence in the certain community.

Figure 4: Layout of possible contact of population and UXO.

Figure 4 shows the territory that presents possible contact of population and mine explosive ordnance. Obtained area, presenting an “objective estimated risk” for domestic populations, is calculated by multiplying the population that is within, or on, the border of suspected mine areas with average population density on the study area. Mine accident risk on the present infrastructure is calculated indirectly as well (i.e., around digitalised installation infrastructure, a 100-m both-sided buffer is determined, and after that by implementation of geoprocessing function an intersection area of minefields and infrastructure installation is determined.) In a similar manner, for the mine-contaminated areas of each of the 11 analysed communities, the values of estimated risk for other criteria are evaluated (roads, agriculture areas, forests, parks of nature, etc.).

Municipalities Suspicious and Mine- Contaminated Area (km2) Population on Mine Contaminated Area Number of Recognized Mines
Dvor 21.14 120 161
Glina 24.07 1140 144
Dubica 0.43 0 109
Petrinja 6.94 0 -
Sisak 0.58 0 -
Sunja 11.91 400 438
Topusko 10.74 1600 1271
Kostajnica 77.78 1242 7557
Jasenovac 24.22 2128 10241
Novska 39.10 3730 4962
Gvozd 1.99 50 -

Table 1: Landmine problem in each municipality.

During multicriteria analysis for each of the criteria, the weights were assigned by stakeholders involved in the decision process. Namely, it is important to involve representatives of social and political associations from the communities’ territory, which are included in the priority ranking, in order to obtain results that would be accepted by them as optimal ones.

For the numerical part of multicriteria analysis, namely the PROMETHEE and GAIA methods, “Decision Lab 2000” is used. It is the commercial name of software distributed by Visual Decision from Canada. Contemporary architecture of this software, based on the Decision Support System (DSS), enables wide support in the decision-making processes.

A large portion of the information, most of which is possible to visualise (graphs, various coloured diagrams), gives complete insight to the decision-maker into the problem characteristics and possible results of various problem-solving scenarios.

Table 2 presents results of the numerical analysis by PROMETHEE method, i.e., evaluated ranks that presented optimal policies for risk reduction in mine-contaminated areas (presented results are not final optimal solutions). Achieved synthetic parameter “Phi” presents valorisation of total risk based on defined criteria and weighting coefficients. Table 2 shows that community Sunja is ranked first and represents demining priority, because the total risk of 0.5364 dominates the second-ranked community Petrinja with Phi value of 0.3077. Following the ranks of other communities to the last one, community Gvozd has a negative priority value of Phi -0.2397.

Municipalities Phi Plus Phi Minus Phi Net Ranking
Dvor 0.1830 0.1537  0.0293 4
Glina 0.1078 0.1470 -0.0391 5
Dubica 0.0657 0.1797 -0.1140 7
Petrinja 0.3888 0.0810  0.3077 2
Sisak 0.1043 0.1558 -0.0514 6
Sunja 0.5803 0.0439  0.5364 1
Topusko 0.0572 0.2088 -0.1516 9
Kostajnica 0.0193 0.2206 -0.2014 10
Jasenovac 0.1903 0.1356  0.0547 3
Novska 0.0589 0.1896 -0.1308 8
Gvozd 0.0003 0.2401 -0.2397 11

Table 2: Ranked municipalities.

Figure 5 shows the layout of the relations between criteria obtained by GAIA software, namely by applying of principal component analysis for Phi values for each criteria. Insight into the criteria relations is important for understanding problems and recognising correlations between different UXO risk parameters.

Figure 5: Geometrical presentation of the relations between criteria is shown by the GAIA method.

Conclusion

Multicriteria analysis of humanitarian demining action in the Sisacko-Moslavacka County pointed out many methodological, social and political advantages to this approach to the real and complex problem. The procedure of choosing the optimal policies for risk reduction in mine-contaminated areas when using GIS analysis and multicriteria analysis demands collaboration of social and political authorities and practically involves all interested parties, which are numerous in the humanitarian demining problem. Namely, between small farmers whose backyards are contaminated and county and community councils, forums, and representatives, there are several levels that are directly or indirectly exposed to the mine accident risks. Most of them expect that their problem should be treated as the first priority, so their involvement in the decision process lowers tensions and partly removes frustrations because of problem solving prolongation. On the other side, however, insight into the priority evaluation procedure, in regards to the objectively evaluated risk level, creates a trust climate and a firm standpoint that priorities are evaluated properly because they are involved in the process. Transparency of the available data that are the base for analysis is important because everyone can check whether “his” parameters are correctly evaluated. Therefore, Decision Lab 2000 software has various options for post-analysis and “what-if” simulation estimation (such as the “Walking Weights” option) in order to eliminate subjectivity that is always present during risk evaluation.

References

  1. Brans, J.P. and Mareschal, B. (1994) ‘PROMCALC & GAIA: A new decision support system for multicriteria decision aid,’ Decision Support Systems, 12, pp.297–310.
  2. Buzolić J., Mladineo N., Knezić S., GIS based fire protection management for fire risk zones, Management Information Systems incorporating GIS and Remote Sensing, Brebbia (ed.). Wessex Institute of Technology Press, Southampton, UK, 2000.
  3. Fiedler T., Bajic M., Gorseta D.: GIS for demining activities in Croatia, Joint Research Centre–European Commission, Proceedings Demining Technologies International Exhibition, Workshop and Training Seminars, 29 September–1 October 1998, Ispra, Italy, pp.187–193.
  4. Mladineo, N., Lozić, I., Knezić, S., Mlinarić, D., Radica, T., “An evaluation of multicriterional analysis for DSS in public policy decision,” European Journal of Operational Research, North–Holland, No. 61, pp. 219–229.,1992.
  5. Zionts, S., Ed. (1989) MCDM and Risk Analysis, Computer and Systems Sciences, Springer Verlag.
  6. http://www.visualdecision.com/PROMETHEE

Contact Information

Damir Goršeta, M.Sc.
Croation Mine Action Center (CROMAC)
Ante Kovačića 10 p.p. 8
44 000 Sisak, Croatia
Phone: ++385 44 554-100
Fax: ++385 44 554-111
E-mail: hcr@hcr.hr
Website: http://www.hcr.hr