Repeated-drive adaptive feedback identification of network topologies

Area of Science:

• Complex networks
• Network science
• Dynamical systems

Background:

• Identifying network topology from dynamics is a challenging inverse problem.
• Node synchronization complicates inferring network structure from dynamical data.
• Short-time dynamical data presents difficulties for accurate topology identification.

Purpose of the Study:

• To present an efficient method for revealing network connectivity from short-time dynamics.
• To address the challenge of inferring network topology in the presence of node synchronization.
• To develop a scheme capable of determining the adjacency matrix from transient dynamical data.

Main Methods:

• A repeated-drive adaptive feedback scheme was employed.
• Short asynchronous transient data was used as a repeated drive.
• The method was validated for both synchronous and asynchronous network cases.

Main Results:

• The proposed scheme successfully determined the adjacency matrix using short asynchronous transient data.
• The method is efficient even with global or local synchronization.
• Detection speed can be optimized by adjusting time-series segment length and coupling strength.

Conclusions:

• The repeated-drive adaptive feedback scheme effectively identifies complex network topology from short-time dynamics.
• This method overcomes challenges posed by node synchronization.
• The scheme offers an efficient and optimizable approach for network structure inference.