Developing an objectives hierarchy for multicriteria decisions on land use options, with a case study of biodiversity conservation and forestry production from Papua, Indonesia

A framework was developed for the construction of an objectives hierarchy for multicriteria decisions in land use planning. The process began through identification of fundamental objectives; these were iteratively decomposed into a hierarchy of subobjectives until a level was reached at which subobjectives had measurable attributes. Values were derived for attributes through a variety of methods and weights assigned to objectives through preference elicitation. The framework assumed that the objectives could be incorporated into a linear value function; this required attributes to satisfy preference and difference independence conditions. Strategies were developed to address typical features that distinguish land use decisions from many other multicriteria decisions. The methodology was illustrated with a case study of land use planning in a forestry concession in the Merauke region of Papua Province, Indonesia. The problem involved severe hard constraints; the analysis showed how these can be accommodated within the framework. Results integrated interests and preferences of a diverse set of stakeholders (resident peoples, developers, and conservation professionals) and were intended for implementation. This methodology is extendible to other land use problems.

Daugiatikslių sprendimo priėmimo metodų jautrumo analizė taikant Monte Karlo modeliavimą

Daugiakriteriams sprendimo priėmimo uždaviniams spręsti taikant kiekybinius daugiatikslius sprendimo priėmimo metodus, dažniausiai neatsižvelgiama į galimas pradinių duomenų – rodiklių reikšmių paklaidas. Todėl neaišku, ar gautas sprendimas (alternatyvų išrikiavimas pagal racionalumą) nesikeis, jeigu keistųsi rodiklių reikšmės (± n %, n< 50 %). Šiame straipsnyje, taikant Monte Karlo modeliavimą, atliekama kiekybinių daugiatikslių sprendimo priėmimo metodų – TOPISIS, SAW ir COPRAS – jautrumo analizė. Jautrumo analizės pagrindas yra rodiklių reikšmių pasiskirstymas. Darbo tikslas – sukurti jautrumo analizės modelį, kuriuo būtų galima nustatyti daugiatikslių metodų jautrumą rodiklių reikšmių pasiskirstymo dėsnio atžvilgiu ir įvertinti šiais metodais gauto sprendimo patikimumą. Pateikto jautrumo analizės modelio etapai: 1) rodiklių pseudoatsitiktinių reikšmių generavimas pagal tolygųjį ir normalųjį dėsnius; 2) alternatyvų racionalumų vertinimas nagrinėjamais metodais, naudojant generuotas sprendimų matricas; 3) metodų jautrumo ir gauto sprendimo patikimumo vertinimas. Modeliuojant sprendimo paramos sistemas, atliekant skaičiavimus sprendimo priėmimo metodais, rekomenduojama atsižvelgti į planuojamų taikyti metodų jautrumą, kad padidėtų sprendimo patikimumas.The Sensitivity Analysis of the Multiattribute Decision Making Methods by Monte Carlo SimulationRūta Simanavičienė, Leonas Ustinovičius SummaryThe quantitative multicriteria decision making methods are used for decision making, but the biases of the values of the attributes are ignored often. Therefore it is not clear, if the final decision were changed, when changing the values of attributes. There are scientific research papers intended for the sensitivity analysis of multicriteria decisions, according to the significances of attributes. This paper analyses the sensitivity of quantitative multiattribute decision making TOPSIS – Technique for Order Preference by Similarity to Ideal Solution, SAW – Simple Additive Weighting, COPRAS – COmplex PRoportional ASsessment methods, according to attribute distribution. On this basis of Monte Carlo modeling (simulation) the generation of pseudorandom attribute values is being made, according to uniform and normal distribution. Using generated data the efficiency of alternatives is carried out by the above methods. As a result the sensitivity of the methods and the reliability of final decision are presented. When multiattribute decision making methods are used on the modeling decision support systems it is recommended to consider the sensitivity of these methods.