Abstract: Achieving agreement is a fundamental problem in the study of multi-agent systems. A basic understanding is that sufficient connectivity of the network will lead to the state agreement of multi-agent systems. In the commonly considered networks, the exchanged information is assumed to be local and the relationships between different agents are considered to be cooperative. However, in certain, or even most of network setting, there are also global information that can be used for agreement seeking and antagonistic interactions that may affect state evolution. In this talk, we try to understand what the role of global information is in the classical multi-agent model and how antagonistic interactions will influence state evolution of the network. In particular, we validate that the global information is useful for agreement seeking in the sense of relaxing the fundamental convexity condition of the classical multi-agent model and accelerating the convergence speed. In addition, we reveal that both cooperative and antagonistic interactions contribute to state convergence of cooperative-antagonistic network and strong connectivity, instead of quasi-strong connectivity, is critical.

Short bio: Ziyang Meng received his Ph. D. degree from Tsinghua University, Beijing, China, in July 2010. He was a visiting Ph.D. student at Utah State University, Logan, USA from Sept. 2008 to Sept. 2009 and a postdoctoral researcher at, respectively, Shanghai Jiao Tong University, Shanghai, China from Oct. 2010 to July 2012 and KTH Royal Institute of Technology, Stockholm, Sweden from Aug. 2012 to Aug. 2014. He has been with Technische Universitat Munchen, Germany, as a Humboldt research fellow since Sept. 2014. His research interests include multi-agent systems, distributed/decentralized control, spacecraft systems, and nonlinear systems.