摘要: The complexity and uncertainty of objective thing the fuzziness human thought result in decision making with linguistic information a wide variety practical problems, such as personnel evaluation, military system performance online auctions, supply chain management, venture capital, medical diagnostics. In realistic approach may be to use assessments instead numerical values by means variables, that is, variables whose are not numbers but words or sentences natural artificial language (Fan Wang, 2003; 2004; Fan et al., 2002; Herrera Herrera-Viedma, 2000a; 2000b; 1997; Martinez, 2001a; 2001b; Verdegay, 1993; 2005; 2000; 1996a; 1996b; 1995; 2001; Herrera-Viedma Peis, Wang Chuu, Xu, 2010; 2009a; 2009b; 2009c; 2008; 2007a; 2007b; 2007c; 2007d; 2007e; 2006a; 2006b; 2006c; 2006d; 2006e; 2006f; 2006g; 2006h; 2006i; 2005a; 2005b; 2005c; 2005d; 2004a; 2004b; 2004c; 2004d; 2004e; 2004f; Xu Da, Zadeh Kacprzyk, 1999a; 1999b). This arise for different reasons (Chen Hwang, 1992): ① unquantifiable due its nature; ② precise quantitative stated because either it is unavailable cost computation too high an “approximate value” tolerated. For example, when evaluating “comfort” “design” car, labels like “good”, “fair” “poor” usually used, car’s speed, “very fast”, “fast” “slow” can used (Bordogna Levrat 1997).