Coarse-graining Complex Networks for Control Equivalence
Authors:
- Daniele TollerDTDaniele Toller
- Mirco TribastoneMTMirco Tribastone
- Max TschaikowskiMTMax Tschaikowski
- Andrea VandinAVAndrea Vandin
Preprint
Abstract
The ability to control complex networks is of crucial importance across a wide range of applications in natural and engineering sciences. However, issues of both theoretical and numerical nature introduce fundamental limitations to controlling large-scale networks. In this paper, we cope with this problem by introducing a coarse-graining algorithm. It leads to an aggregated network which satisfies control equivalence, i.e., such that the optimal control values for the original network can be exactly recovered from those of the aggregated one. The algorithm is based on a partition refinement method originally devised for systems of ordinary differential equations, here extended and applied to linear dynamics on complex networks. Using a number of benchmarks from the literature we show considerable reductions across a variety of networks from biology, ecology, engineering, and social sciences.Keywords
All Related Versions
This publication has 20 references indexed in Scilit:
- Symbolic computation of differential equivalencesTheoretical Computer Science, 2019
- Maximal aggregation of polynomial dynamical systemsProceedings of the National Academy of Sciences, 2017
- Energy scaling of targeted optimal control of complex networksNature Communications, 2017
- Physical controllability of complex networksScientific Reports, 2017
- Target control of complex networksNature Communications, 2014
- McKay’s canonical graph labeling algorithm2009
- Approximation of Large-Scale Dynamical Systems: An OverviewIFAC Proceedings Volumes, 2004
- Abstractions of Hamiltonian control systemsAutomatica, 2003
- Optimal state-space lumping in Markov chainsInformation Processing Letters, 2003
- The explicit linear quadratic regulator for constrained systemsAutomatica, 2002