Research

From SYSOS

Revision as of 09:43, 10 October 2022 by Antonis (talk | contribs) (Created page with "== Research Interests == Our group develops tools and algorithms for the analysis and design of feedback control in biological and technological systems and applies them in a...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Research Interests

Our group develops tools and algorithms for the analysis and design of feedback control in biological and technological systems and applies them in a range of areas, from Synthetic Biology to fluid mechanics. In particular, we have been developing theory to understand how nonlinear networked systems (System of Systems) operate how to design control laws for them, using computational tools based on the Sum of Squares (SOS) decomposition and Semidefinite Programming (SDP). We have been applying this theory to understand and (re)design biological systems (Systems and Synthetic Biology) but also to analyse and design control laws for fluid flows, robust synchronization, multi-agent system consensus, smart power networks, the robustness of (deep) neural network controllers and congestion control for the Internet. Here is a short summary of the areas we work on - please follow the links below for more details.

  • Nonlinear systems analysis, Sum of Squares optimization and Lyapunov techniques. In this spirit, we have developed methods for stability analysis of systems described by Ordinary, Delay and Partial Differential Equations, as well as switched/hybrid systems. Of current interest is how to scale these tools to handle much more complicated and richer system descriptions, of higher dimension (see also point below, on scaling SDP/SOS methods).
  • Systems and Synthetic Biology. We collaborate closely with groups in Engineering Science (Professors Wei Huang and Harrison Steel) and Biology (Professor Philip Poole) and other institutions for understanding biological pathways through mathematical modelling using experimental data and designing new experiments for model invalidation (Systems Biology), as well as proposing and implementing redesigns for existing biological systems for improved performance (Synthetic Biology).
  • Software Development (SOSTOOLS). This can be found here.
  • Large-scale Networked Systems analysis with communication and structural constraints. Examples in this field come from synchronization phenomena in oscillator networks, Network Congestion Control for the Internet and Power System analysis/Smart grid. We are also looking at multi-agent systems consensus under communication and structural constraints such as the effect of time delays and switching topologies.
  • Fluid mechanics and Heat transfer, from a control perspective. We are taking a control engineering approach to understand questions in Hydrodynamic Stability and mechanisms for background noise energy amplification and subsequent reduction. We are also using a Sum of Squares approach for the stability analysis and control design of systems described by Partial Differential Equations.
  • Complexity reduction of networked systems. We are investigating methods based on model decomposition and reduction to facilitate computational analysis and identification of large-scale and nonlinear systems. Systems of interest include biochemical reaction networks, power systems and consensus networks.
  • Scaling Sum of Squares and Semidefinite Programming. We are using sparsity- and symmetry-exploiting algorithms for improving the scalability of semidefinite programming, including the development of software (such as CDCS) to implement relevant operator-splitting algorithms.
  • Robust analysis and design of (deep) neural network controllers. We are using tools from traditional and modern control theory for the analysis and design of (deep) neural network controllers used in (deep) reinforcement learning.