Scalable Design of Heterogeneous Networks

Abstract

A systematic approach to the analysis and design of a class of large dynamical systems is presented. The approach allows decentralised control laws to be designed independently using only local subsystem models. Design can be conducted using standard techniques, including loopshaping based on Nyquist and Popov plots, H1 methods, and µ-synthesis procedures. The approach is applied to a range of network models, including those for consensus, congestion control, electrical power systems, and distributed optimisation algorithms subject to delays.

H-Infinity Optimal Control for Systems With a Bottleneck Frequency

“…However, in [11]- [13] methods for efficient computation of distributed controllers for so called positively dominated systems are presented. Local stability conditions are demonstrated in [14], [15] that are independent of the system's network and size. Similarly, [16] presents a scalable stability criterion for certain interconnected systems.…”

CMOS Implementation of ANNs Based on Analog Optimization of N-Dimensional Objective Functions

“…3. A dynamic system implemented with basic differential equations exhibits more robustness to certain parameter variation [9,10]. 4.…”

Topology-Independent Robust Stability Conditions for Uncertain MIMO Networks

“…In the frequency domain, robust stability conditions for interconnections of either SISO (single-input and single-output) or MIMO (multiple-input and multiple-output) linear systems were provided in [9], [10], [11], [12], [13] adopting Nyquisttype approaches and in [7], [8], [14], [15] using the generalised frequency variable framework; also, based on Integral Quadratic Constraints, [16], [17] provided scalable conditions that can be tested locally and used for control design [18]. Frequency-domain conditions for topology-independent robust stability were derived in [19], [20] for nominally homogeneous SISO systems and in [21] for homogeneous MIMO systems.…”