Further Improvements on Non-Negative Edge Consensus of Networked Systems

Area of Science:

• Control Theory
• Networked Systems
• Graph Theory

Background:

• Addressing the non-negative edge consensus problem in positive networked systems is crucial for stability and performance.
• Existing methods often have conservative criteria or lack efficient computational approaches.

Purpose of the Study:

• To develop novel and less conservative criteria for achieving non-negative edge consensus in positive networked systems.
• To introduce both linear and semidefinite programming (SDP) based approaches for controller design.

Main Methods:

• Utilizing state-feedback protocols for systems with undirected graphs.
• Developing improved criteria for consequentiality and non-negativity, enabling linear programming (LP) solutions.
• Deriving necessary and sufficient criteria for an SDP approach, including improved bounds for Laplacian and degree matrices.
• Introducing slack matrix variables to separate system, controller, and Lyapunov matrices for easier parameterization.

Main Results:

• Significantly improved criteria for consequentiality and non-negativity, leading to an LP approach.
• Necessary and sufficient criteria for an SDP approach, offering a less conservative alternative.
• A sufficient condition for consensus that only requires the number of edges, independent of network topology.
• Demonstrated reduced conservativeness compared to existing methods through analytical and numerical comparisons.

Conclusions:

• The proposed LP and SDP approaches provide effective and less conservative solutions for the non-negative edge consensus problem.
• The developed criteria enhance the understanding and control of positive networked systems.
• The findings offer practical advancements for designing controllers in networked systems.