ON THE FUNDAMENTAL CHARACTERISTICS OF COMPLEX NETWORK WITH MULTI-AGENT CONSTITUENTS
Chun-Lin Yang1; C. Steve Suh1
1 Mechanical Engineering Department
Resumo
Multi-agent systems incorporating the concept of complex network are studied to gain insight into the dynamics that governs the collective behavior of the network. A network system along with its behavior can be defined by 3 primary network parameters; namely, Degree Distribution, Cluster Coefficient, and Average Path Length. Real-world multi-agent systems such as flocking birds, swarm fish, drone fleets, and human teams are networks whose behaviors are both nonlinear and nonstationary. Since the three network parameters defining the relationship between the nodes (agents) of the network are time-dependent, nodes consensus and network synchronization would be hard to attain. Control methodologies considered effective in enabling the consensus control of any complex network with multi-agents and time-delays, therefore, needs to address the essential aspects of nonlinearity and nonstationarity to be viable. In this presentation, a general framework for developing such a consensus control methodology is discussed which also demonstrates applicability to time-delayed complex networks. The significance of nonlinearity and non-stationarity of real-world networks such as multi-agent systems to consensus control is considered. The correlations of Degree Distribution and node number to the number of edges are explored. Two sets of experimental results are provided. Finally, a general framework for developing such a consensus control methodology is discussed.
Palavras-chave: Complex Network; Nonlinear Dynamics and Complex Systems; Synchronization in Nonlinear System; Multi-agent Systems