Certifying Collective Reasoning in Multi-Agent Systems via Koopman Spectral Analysis

DOI10.48550/arXiv.2608.05956MaRDI QIDQ7284740arXivFDO

Authors Indrakshi Dey, Nuzhat Khan

Publication date 6 August 2026

Published in arXiv

Tldr We analyze multi-agent systems as nonlinear dynamical systems, leveraging Koopman operator theory to certify convergence timescales, factional structure, and decision-explaining coordinates from interaction traces.

Full work available at URL https://doi.org/10.48550/arXiv.2608.05956

arXiv classification

cs.MAeess.SY

Cites work

• Consensus Problems in Networks of Agents With Switching Topology and Time-Delays
• Data-driven approximation of the Koopman generator: model reduction, system identification, and control
• Diffusion maps
• Dynamic mode decomposition and its variants
• Dynamic mode decomposition of numerical and experimental data
• Extended dynamic mode decomposition with dictionary learning: a data-driven adaptive spectral decomposition of the Koopman operator
• Hamiltonian systems and transformations in Hilbert space.
• Modern Koopman theory for dynamical systems
• On convergence of extended dynamic mode decomposition to the Koopman operator
• Reaching a Consensus
• Spectral analysis of nonlinear flows
• Spectral properties of dynamical systems, model reduction and decompositions

This page was built for publication: Certifying Collective Reasoning in Multi-Agent Systems via Koopman Spectral Analysis

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7284740)