polska wersja

News

About (short description)

With numerous parties in the parliament it is not easy to assess the power of each one of them. Typically, a party with more than a half of all seats obviously has all the power as it can pass any bill. Losing the majority even by a single vote changes the situation diametrically - decisions have to be made by coalitions, and the power becomes a complex function of the number of seats held by each party.

But mathematicians have a special tool designed exactly for such analysis - weighted voting games. In this very simple model each player (in our example representing a political party) has a number of votes and a decision is made if the number of players' votes in support of it exceeds a certain threshold. From the model, there is only one step to calculate the power - to this end, power indices has been proposed, that consider how often each player could provide the swing votes that would convert the losing side into the winning side.

As most theoretical models, weighted voting games are based on various simplifying assumptions. In particular, the model assumes that any player is always ready to cooperate with all others. In reality, however, this is usually not the case, due to sympathies and antipathies between political parties.

The key objective of our project is to propose a more realistic model of weighted voting games which addresses the aforementioned problems and to develop as efficient as possible techniques to calculate power indices within this model.

Members


Oskar Skibski
Principal Investigator

University of Warsaw, PL


Collaborators


Talal Rahwan
Collaborator

Masdar Institute, UAE

Yuko Sakurai
Collaborator

AIST Tokyo, JP

Takamasa Suzuki
Collaborator

Gifu University, JP

Michael Wooldridge
Collaborator

University of Oxford, UK

Makoto Yokoo
Collaborator

Kyushu University, JP


Publications (so far)