Analysis and control of large scale complex networks
Lead Research Organisation:
University of Cambridge
Department Name: Engineering
Abstract
The project is associated with modelling, analysis and control of large scale complex networks of interacting dynamical systems, with a main emphasis on how the underlying noise inevitably present in the interactions can affect the system functionality and emergent behaviour. In particular, the role of noise in biochemical reaction networks will be studied first investigating how the intrinsic stochasticity of biochemical reactions can be suppressed by means of feedback. Noise is ubiquitous within the cell due to the spontaneous birth and deaths of individual molecules. This noise could drive metabolites away from their desired concentrations or sometimes be beneficial for diversity. The discreteness of these molecular events in conjunction with the randomness in the time between them introduces challenges in the analysis relative to more conventional methods in control theory. Throughout the project analytical methods for quantifying such noise will be developed as well as hard limits for its suppression. The effect of stochasticity in other type of networks will then be explored such as power networks with stochasticity in the supply/demand and the communicating signals.
Organisations
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/N509620/1 | 01/10/2016 | 30/09/2022 | |||
| 2107279 | Studentship | EP/N509620/1 | 01/10/2018 | 30/09/2021 | Giovanni Pugliese Carratelli |
| Description | In the first part of the project we have focused on the modelling and the analysis of noise within biochemical reaction networks. We have devised an analytical method to quantify the noise for a class of biochemical networks. In particular, we have considered a setting where the growth rate of a molecule is non-linear and influenced by the abundance of another molecule. Classical tools to compute the noise in such a setting include approximate methods or computationally intensive simulations. The analytical tool we have proposed, computes a hard bound on the level of noise and is important from both an analytical and computational point of view. The problem addressed leads also to a number of challenging questions. Specifically, generalising the setting to an arbitrary number of molecules is an interesting but non-trivial problem and could be part of future work. In the second part of the project, also in light of the recent pandemic, we have turned our attention to methods for the control of stochastic population dynamics. Properties of optimal control policies have been investigated, and these have been analyzed when used in conjunction with optimal real-time filters. This investigation is still ongoing and preliminary results will be presented at an upcoming conference. |
| Exploitation Route | The framework may be of use to the quantitative biologists who are seeking for a rapid way to evaluate the level of noise in simple bio chemical networks. Moreover, an area of research could be to analyse if these methods can be extended to handle to settings where feedback is present which may be an area of interest for control engineers and mathematical biologist. |
| Sectors | Other |
| Description | While the first part of the project, which has focused on noise within biological networks, has not had non-academic impact thus far, the current research regarding mitigation strategies for epidemics will most likely have an impact on decision making and benefit society as a whole. |