Multiagent dynamics and collective decisions on signed networks

Signed networks appear for instance when modeling multiagent interactions which are both of cooperative and antagonistic nature (e.g. social networks).

The multiagent dynamics that can occur on such signed networks are normally richer and much more complex than those possible on positive (i.e., purely cooperative) graphs, and are currently not well understood.

The investigation of dynamics on signed networks is the overarching goal of this proposal.More specifically, the aim is to explore several research directions which can be schematically described as follows: (1):  Investigate the structure and stability of signed Laplacians on signed networks, for applications like DC microgrid small angle disturbance analysis; (2): Understand how antagonism can affect a collective decision-making process on multiagent networks, using tools such as bifurcation analysis; (3): Investigate opinion dynamics of unanimity, polarization and similar emerging behaviors on social networks characterized by antagonism; (4) Develop distributed measures of trust/mistrust in multiagent networks, and algorithms to propagate these trust/mistrust measures in e.g. mobile communication networks.For these research topics, the plan is to develop theoretical methods needed for the analysis, but also algorithms able to apply these instruments to concrete datasets.