Voting is the process of choosing a "best" alternative in view of the preferences of a group of voters over a set of alternatives. In this talk I will present recent work based on the following simple idea: todetermine the winner of an election, first identify the "most representative" voter in the group and then implement the will of that voter. I will explore this idea in the context of binary aggregation, where each voter expresses yes/no choices regarding a number of possibly correlated issues and we are then asked to decide on a collective choice that accurately reflects the views of the group. We will see that certain rules based on our idea, namely the average-voter rule and the majority-voter rule, have surprisingly good properties. In particular, they allow us to approximate the theoretically attractive but practically intractable distance-based rule up to a small constant factor. I will start the talk with a short general introduction to the field of computational social choice. No specific background in voting theory will be required to follow the presentation.
Institute for Logic, Language & Computation (ILLC)
University of Amsterdam