Using political science voting models to determine weightings in multi-objective decision problems

In this paper we present an approach to the coordination of complex systems operating in a multi-objective domain, which is an extention of the dynamic weighted aggregation method where the weightings are influenced by the decision makers themselves. This is implemented through an information theoretic voting model, taking its inspiration from the political science quantitative models of elections. The example system presented shows this coordination method, where voters vote for the parties (and hence the weightings) corresponding to their preferences, and parties update their policy (weighting) positions learning from the voters.