Nonlinear Opinion Dynamics on the Sphere for Distributed Multi-Agent Systems
Author(s)
Zhang, Ziqiao
Advisor(s)
Editor(s)
Collections
Supplementary to:
Permanent Link
Abstract
Multi-agent systems have been widely studied with various potential real-world applications. Much of the inspiration for this field comes from natural collective behaviors, such as fish schooling, ant colonies, flashing fireflies, etc. When a large swarm of individuals are performing complicated tasks, group-level decision-making is critical to successful task completion based on individual-level information collection and communication. During this process, individuals form various opinions, exchange opinions with their neighbors, and gradually reach a consensus or dissensus, which leads to a group-level decision. Thus, studying the opinion dynamics of multi-agent systems is crucial for understanding the decision-making process from both group and individual perspectives.
The objective of this dissertation is to explore and develop models of opinion dynamics for distributed multi-agent systems, focusing on the diverse behaviors of consensus and dissensus under various interaction rules in unsigned graphs. A novel aspect of this dissertation is the modeling of opinion states as unit-length vectors on a sphere, representing unique measures of expressed opinions. The evolution of these state vectors illustrates the change in opinions of each agent, influenced by neighboring opinions.
The dissertation's contributions are classified into theoretical work and practical applications. The theoretical contributions establish foundational models for understanding rich opinion behaviors, including consensus and various forms of dissensus. These models are significant not only for describing individual and group behaviors in social networks, but also for explaining different communication methods when agents interact and exchange information.
On the application front, the models are applied to multi-robot task allocation. In these contexts, opinion dynamics facilitate upper-level decision-making processes. The research thus bridges theoretical insights and practical implementations, enhancing the understanding and utility of opinion dynamics in complex systems.
This work bridges the gap between theoretical models and applications of opinion dynamics in distributed systems. By developing and applying innovative models, the research contributes to a deeper understanding of how opinions evolve and influence decision-making processes in complex, multi-agent environments. This work not only advances the field of opinion dynamics but also provides valuable insights and tools for practical implementations in various technological and social domains.
Sponsor
Date
2024-12-05
Extent
Resource Type
Text
Resource Subtype
Dissertation (PhD)