Normalizavimo metodų pasirinkimo įtaka priimant sprendimus statyboje pagal lošimų teoriją

Abstract

Summary In decision support systems (DSS) are commonly used as several multicriteria decision making value normalization methods. Their importance in decision making is high enough and must be analyzed comprehensively. The first step is the value transformation to a general parameter system after values are normalized. In this article nine value normalization methods are given (Tables 2 and 3). According to the problem objective tenders choice, their criteria and condition (Table 1), all value normalization methods are used and the analysis is made. Normalized value data sheets are given with a graphical view and final ranking (Table 6). The conclusions are the following: We can divide all value normalization methods in to three groups: I group – methods No 1, 3, 4, 5 and 6 (conformation >70 %); II group – methods No 2 and 9 (conformation ~60 %); III group – methods No 7 and 8 (conformation ~70 %). The normalized values of I group methods are multiplied or are the same (No 3 and 5). In order to prevent the value concentration and wrong final decision it is better to evade the usage of methods No 1 and 6. I group methods are close to ideal, but have the main disadvantage, they exclude inner alternatives from comparison availability list, only the best and the worst values are given. Because normalized values of III group methods are averagely concentrated and distributed, it is better to use them. The final conclusion: the value normalization methods used in multicriteria decision making and DSS have extreme influence on the final ranking of alternatives and the final decision.