Majority decisions based on difference of votes (Q5943174)

For a finite number of voters choosing between two alternatives, \(x\) and \(y\), the authors introduce the ideas of \(M_k\) and \(M^l_k\)-majorities, where the parameter \(k\) measures the difference in preference by the voters between the two alternatives and \(l\) acts as a threshold so that some minimum number of votes is required to choose \(x\) over \(y\), as well as some difference in votes. The threshold reflects the authors concern with situations where there is a large amount of voter abstention or, equivalently, indifference. Both the \(M_k\) and \(M_k^l\) majorities are neutral, monotonic and unanimous. The \(M_k\) and \(M_k^l\) majorities are defined within the context of crisp preferences. In order to extend these majorities to the case where voters have fuzzy preferences, the authors introduce the crucial notion of an \(\alpha\)-discrete aggregation function (\(\alpha\)-DAF). Here, the \(\alpha\) parameter, \(\alpha \in [1/2, 1)\), mediates between the fuzzy input and crisp output. The authors then show that \(M_k\) majorities can be extended (non-uniquely) to fuzzy preferences via an associated quasiarithmetic mean. Similarly, the \(M_k^l\) majorities can be extended to fuzzy preferences via associated ordered weighted average operators.