Page 216 - Materials Chemistry, Second Edition

214 10. Advancing life cycle sustainability assessment using multiple criteria decision making

alternative performing better in at least one of selected attributes and performs equally in
other attributes. In the case of a large number of alternatives in decision making, identifying
and utilizing only the nondominant alternatives is essential and significantly decreases the
efforts required for finding a feasible alternative (Calpine and Golding, 1976). The set of
nondominated alternatives is also known as “Pareto-optimal.” Removing dominated alterna-
tives from the alternatives is optional; however, it increases unnecessary noise in the overall
decision-making process and, considering limited processing capability of decision-makers,
develops a condition which is ex ante worse. Literature suggests that presence of dominated
alternatives in the decision matrix can cause asymmetric dominance effect or attraction effect
(Huber et al., 1982), compromise effect (Simonson, 1989), and similarity effect (Tversky, 1972),
thus influencing the decision process. Additionally, according to Montgomery and Will en
(1999), the decision maker’s tendency of justifying or protecting their selected alternative
adds uncertainty. Therefore, initial development of Pareto-optimal set of alternatives is most
desirable.
The process of identifying nondominant alternatives as an introductory refining process is
a well-utilized concept throughout the literature. Hwang and Yoon (1981) suggested identi-
fication of the nondominated alternatives is tricky and, therefore, suggested some methods
like “dominance,” “permutation method,” and “ELECTRE,” which could help in the process
of identification. Dominance method successively compares two alternatives at a time and
deletes the dominated alternative in each step. Whereas, in the permutation method, first,
attempts are made to identify the best ordering of ranking and then identify the dominated
alternatives. ELECTRE method makes a pairwise comparison of alternatives and the weights
help in supporting or denying the dominance relationship among alternatives. But, as men-
tioned in Section 10.5.1, a distance-based method like TOPSIS is more appropriate than
relative utility-based method like LWS; therefore, for distance-based methods with internal
reference data, Kalbar et al. (2017b) recommend choosing Hasse diagram technique as the
most appropriate technique for identifying dominating alternatives. The method is appropri-
ate due to its simplicity in a pictorial representation of relationships.

10.5.4 Consistency of inputs on indicators

The decision-maker provides information or inputs on indicators related to the decision
problem. However, measuring consistency in the provided information is essential as it in-
fluences the results. Consistency can be affected by condition of preferential independence
and dependency (Waas et al., 2014; Figueira et al., 2005). Preferential independence is the con-
dition in which preference of one indicator over another indicator is not affected by any other
existing indicator. However, any MADM problem is rarely free from this condition, as some
of the selected indicators may have some level of interaction (Liou and Tzeng, 2012; Tzeng
and Huang, 2011; Triantaphyllou, 2000). In MADM, there is a general assumption that indi-
cators are independent of each other. Hence, the results of MADM are not certain, due to lack
of understanding in trade-offs among indicators. A detailed approach to check preferential
independence using the trade-off method of preference level of indicators is given in Figueira
et al. (2005). It is recommended that if interdependence exists among a specific set of indica-
tors, then decision-makers should attempt to group the indicators or break indicators with