Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Stability and controllability of multi-agent systems over stochastically-evolving topology
Multi-agent systems constitute an applicable class of dynamical systems wherein multiple individual systems are coupled in various forms, e.g., physical interconnections or coupled objectives. Stability and controllability of such systems are fundamental to understand their complex dynamical behavio
Keywords: Multi-agent systems, stability and controllability, stochastic evolution, random topology
Collective behavior of multi-agent systems is determined by the dynamical behavior of each
individual system as well as the interaction model between the agents that is often modeled by a graph. In many applications of multi-agent systems, the interaction pattern is not fixed and evolves with time where this evolution may follows non-deterministic models. Therefore, classical stability and performance concepts need to be extended to investigate the properties of such special class of multi-agent systems. The main contribution of this project is to consider a (semi-)random coupling topology for the individual systems, such that the interactions between every two agents will occur in a stochastic fashion, i.e. each interaction exists according to some probability distributions. Stability and controllability analysis of such systems require further studies of stochastic dynamical systems and control over time-varying topology. The mentioned model with random topology is of notable application importance in the domain of power systems, network of oscillators, and also recently in neurology.
Collective behavior of multi-agent systems is determined by the dynamical behavior of each individual system as well as the interaction model between the agents that is often modeled by a graph. In many applications of multi-agent systems, the interaction pattern is not fixed and evolves with time where this evolution may follows non-deterministic models. Therefore, classical stability and performance concepts need to be extended to investigate the properties of such special class of multi-agent systems. The main contribution of this project is to consider a (semi-)random coupling topology for the individual systems, such that the interactions between every two agents will occur in a stochastic fashion, i.e. each interaction exists according to some probability distributions. Stability and controllability analysis of such systems require further studies of stochastic dynamical systems and control over time-varying topology. The mentioned model with random topology is of notable application importance in the domain of power systems, network of oscillators, and also recently in neurology.
The goal of this project is to analyze the stability properties of multi-agent systems over (semi-)random topology graphs, meaning the physical interconnections between the agents follow a stochastic pattern over time. Moreover, we are interested in fundamental controllability analysis of such systems in form of proposing a verifiable controllability test.
The goal of this project is to analyze the stability properties of multi-agent systems over (semi-)random topology graphs, meaning the physical interconnections between the agents follow a stochastic pattern over time. Moreover, we are interested in fundamental controllability analysis of such systems in form of proposing a verifiable controllability test.
Dr. Mohammad (Vahid) Mamduhi (mmamduhi@ethz.ch), Prof. John Lygeros (jlygeros@ethz.ch)
Dr. Mohammad (Vahid) Mamduhi (mmamduhi@ethz.ch), Prof. John Lygeros (jlygeros@ethz.ch)