Abstract
A networked multiagent system is a dynamic system that consists of multiple
interconnected subsystems called agents. These agents are interconnected
according to a certain communication topology: the network graph.
An important problem in the theory of multiagent systems is the problem of
robust synchronization. Here, the dynamics of agents are uncertain. While the
dynamics of the individual agents are not known precisely, they are close to a
given nominal dynamics. The goal is to find communication protocols that achieve
synchronization despite this uncertainty. A network is said to be synchronized
if the states of the agents converge to a common trajectory. In the first part
of this thesis, we provide protocols that achieve synchronization robustly for
networks where the agent dynamics are uncertain in the sense that the nominal
dynamics has been perturbed by coprime factor perturbations.
Another wellknown problem is that of model reduction: given a complex,
largescale system, can we find less complex models that accurately approximate
our original system? Applying existing techniques to networks in general leads
to unsatisfying results, because the model reduction step destroys the structure
of the network. In the second part of this thesis, we investigate two different
techniques that preserve the structure of the network. The first technique is
based on clustering. Here, the agents are divided into groups and each group is
represented by a single agent in the reduced network. The second technique
instead reduces the network graph by removing cycles, thus reducing the
complexity of the communication topology.
interconnected subsystems called agents. These agents are interconnected
according to a certain communication topology: the network graph.
An important problem in the theory of multiagent systems is the problem of
robust synchronization. Here, the dynamics of agents are uncertain. While the
dynamics of the individual agents are not known precisely, they are close to a
given nominal dynamics. The goal is to find communication protocols that achieve
synchronization despite this uncertainty. A network is said to be synchronized
if the states of the agents converge to a common trajectory. In the first part
of this thesis, we provide protocols that achieve synchronization robustly for
networks where the agent dynamics are uncertain in the sense that the nominal
dynamics has been perturbed by coprime factor perturbations.
Another wellknown problem is that of model reduction: given a complex,
largescale system, can we find less complex models that accurately approximate
our original system? Applying existing techniques to networks in general leads
to unsatisfying results, because the model reduction step destroys the structure
of the network. In the second part of this thesis, we investigate two different
techniques that preserve the structure of the network. The first technique is
based on clustering. Here, the agents are divided into groups and each group is
represented by a single agent in the reduced network. The second technique
instead reduces the network graph by removing cycles, thus reducing the
complexity of the communication topology.
Original language  English 

Qualification  Doctor of Philosophy 
Awarding Institution 

Supervisors/Advisors 

Award date  17Feb2017 
Place of Publication  [Goningen] 
Publisher  
Print ISBNs  9789036795531 
Electronic ISBNs  9789036795524 
Publication status  Published  2017 