Projectdetails

This project will investigate the problem of aggregating the preferences of a group of individual agents into a collective decision when the individual agents do not have full certainty about the consequences of the decision to be agreed upon. This problem is of immediate scientific interest as it provides a testbed for the integration of two distinct lines of work in Artificial Intelligence that have sought to apply computational methods to issues traditionally studied in mathematical economics: decision making under uncertainty and computational social choice. The problem is also relevant to a number of applications in the ICT domain, for instance, to the challenge of supporting automated consensus finding in social networks. The project has two main objectives: (1) to identify the most important sources of uncertainty that affect an agent's preferences, leading to the design of formal languages to compactly represent preferences and uncertainty in an integrated fashion, and (2) to develop methods for the aggregation of such preferences over uncertain outcomes. The planned programme of work has both a theoretical component, devoted to the rigorous analysis of the properties of the representation languages and interaction protocols devised, and a practical component aimed at developing a test suite for the empirical evaluation of the proposed methods.