arXiv++ Combinatorics

The expected number of pairwise comparisons needed to learn a partial order on n elements is shown to be at least n*n/4-o(n*n), and an algorithm is given that needs only n*n/4+o(n*n) comparisons on average. In addition, the optimal strategy for learning a poset with four elements is presented.

Different types of two- and three-dimensional representations of a finite metric space are studied that focus on the accurate representation of the linear order among the distances rather than their actual values. Lower and upper bounds for representability probabilities are produced by experiments including random generation, a rubber-band algorithm for accuracy optimization, and automatic proof generation. It is proved that both farthest neighbour representations and cluster tree representations always exist in the plane. Moreover, a measure of order accuracy is introduced, and some lower bound on the possible accuracy is proved using some clustering method and a result on maximal cuts in graphs.

Actual individual preferences are neither complete (=total) nor antisymmetric in general, so that at least every quasi-order must be an admissible input to a satisfactory choice rule. It is argued that the traditional notion of ``indifference'' in individual preferences is misleading and should be replaced by `equivalence' and `undecidedness'. In this context, ten types of majority and minority arguments of different strength are studied which lead to social choice rules that accept profiles of arbitrary reflexive relations. These rules are discussed by means of many familiar, and some new conditions, including `immunity from binary arguments'. Moreover, it is proved that every choice function satisfying two weak Condorcet-type conditions can be made both composition-consistent and idempotent, and that all the proposed rules have polynomial time complexity.