Decomposition and Structured Model Reduction for Large Scale Systems Analysis
- đ¤ Speaker: James Anderson, University of Oxford
- đ Date & Time: Thursday 21 January 2016, 14:00 - 15:00
- đ Venue: Cambridge University Engineering Department, LR5
Abstract
In order to analyse or design a controller for a dynamical system one often needs to first reduce the state dimension. In the realm of networked systems this is true even for Linear Time Invariant (LTI) systems. Traditional approaches to model reduction such as truncation and singular perturbation require one to balance the model first, thus introducing a coordinate transformation on the state variable. For networked systems, balancing annihilates the sparsity pattern in the system matrices, and therefore removes the inherent network structure. In this talk I will present a method for structure-preserving model reduction based on balance truncation. The method comes with error bounds and the reduction can be computed using semidefinite programming in the worst case and linear algebra in the best case. The algorithms developed are applied to mechanical and biological systems and shown to outperform existing approaches.
Series This talk is part of the CUED Control Group Seminars series.
Included in Lists
- All Talks (aka the CURE list)
- bld31
- Cambridge Centre for Data-Driven Discovery (C2D3)
- Cambridge talks
- Cambridge University Engineering Department, LR5
- Cambridge University Engineering Department Talks
- Centre for Smart Infrastructure & Construction
- Chris Davis' list
- Computational Continuum Mechanics Group Seminars
- CUED Control Group Seminars
- Featured lists
- Information Engineering Division seminar list
- Interested Talks
- ndk22's list
- ob366-ai4er
- Probabilistic Systems, Information, and Inference Group Seminars
- rp587
- School of Technology
- Trust & Technology Initiative - interesting events
- yk449
Note: Ex-directory lists are not shown.
![[Talks.cam]](/static/images/talkslogosmall.gif)

James Anderson, University of Oxford
Thursday 21 January 2016, 14:00-15:00