Skip to main content

Microscopically-informed active field theories

Objectives

Active field theories are widely used to study collective effects in driven systems at all levels of organisation, allowing instabilities to pattern formation to be identified [1]. Most commonly such theories are phenomenological, built by identifying the dominant terms consistent with symmetries and conservation laws that are present [2,3]. It is difficult in this approach to connect to underlying microscopic processes, such as the rate at which a microscopic constituent (be that a motor protein, a cell or a higher organism) consumes energy. In this project, our aim is to build on previous success [4] in constructing a stochastic field theory from first principles that features multiplicative noise, and where quantitative agreement with a particle-based model is achieved in one spatial dimension with a numerical algorithm that employs non-Gaussian stochastic increments. We aim to extend this approach to more complex models in higher-dimensions, thereby creating new microscopically-informed field-theories for dry active matter. This could be used to predict phase transitions driven by changes in behaviour at the microscopic scale.

Comparison of an agent-based model for flocking and schooling (arrows) with an active field theory (coloured patches). Credit: Richard Blythe
Comparison of an agent-based model for flocking and schooling (arrows) with an active field theory (coloured patches). Credit: Richard Blythe

Activities of the Doctoral Candidate

Following the approach of [4], we will construct active field theories for particles undergoing self-propulsion, alignment, attraction and repulsion. These model bird flocks or fish schools in more than one dimension. We will combine model-building and parallel computing expertise to create new field theories and implement novel numerical integration algorithms on high-end GPUs. This will allow efficient parameter space scanning, validation against particle-based simulations and study of collective effects and phase transitions. Models will also be applied to 3d tracking data of fish schooling collected by the Rome group [6]. There is also an opportunity to model swimming sperm which exhibit collective motion at high speeds, applying to data provided by Dyneval.

Facilities Provided

Access to high performance computing facilities for numerical work, including GPUs.

Employment Contract

The Doctoral Candidate will be employed by the University as a Marie Curie fellow. Pay and other benefits will follow the general Doctoral Network conditions, see the Eligibility page for details.

Period of Doctorate and Funding

No additional allowances will be paid beyond the initial 36 month period funded by the EU. However, no tuition fee will be due if further time is required to write up the thesis.

References

[1] Cates, ME (2022) in Active Matter and Nonequilibrium Statistical Physics (OUP);
[2] Toner, J, & Tu, Y. (1995) Phys Rev Lett 75:4326;
[3] Wittkowsi. R, et al. (2014) Nat Comm 5:4351 (2014);
[4] Ó Laighléis, E, et al. (2018) Phys Rev E 98:062127;
[5] Maggi, C, et al (2021) Soft Matter 17:3807;
[6] sites.google.com/view/claudio-maggi-cnr/projects

About this research project

Host Institution
The University of Edinburgh
PhD Awarding Institution
The University of Edinburgh

Supervision and secondment arrangements

Lead Supervisor
Richard Blythe (The University of Edinburgh)

Secondments

  • 3-month secondments: GPU training, interaction with experimentalists. (With Claudio Maggi, NANOTEC-CNR)
  • 3-month secondment: semantic data management techniques to parameter scanning, linking to experimental data. (With IndiScale GmbH)

Levels of Biological Organisation

Analysis Techniques


Applying for this research project

Applications will be accepted until Wednesday 1st April 2026.

Apply now

Please check eligibility requirements before applying.