Skip to content

Optimal sequential decisions for multivariable spatio-temporal processes

Study level


Faculty/Lead unit

Science and Engineering Faculty

School of Mathematical Sciences

Topic status

We're looking for students to study this topic.


Dr Helen Thompson
Senior Lecturer in Statistics
Division / Faculty
Science and Engineering Faculty
Dr Kate Helmstedt
Lecturer in Operations Research
Division / Faculty
Science and Engineering Faculty

External supervisors

  • Dr Alberto Elfes


As digital technology starts to disrupt the agricultural sector and it becomes feasible to measure spatio-temporal data that characterise crop development, there is a great opportunity for developing and deploying novel mathematical and statistical tools to assist with the spatio-temporal management of agricultural processes.

This project seeks to develop novel tools for the management of crop growth and yield.

We particularly encourage applications from those traditionally underrepresented in the mathematical sciences including women, racial minorities, those from low socio-economic backgrounds, people with disabilities, and those with English as a second language. The university offers additional support for those who need it, including language assistance and flexible working arrangements.

Research activities

The first stage of the project will be to formulate the sequential decision problems that will assist farmers in planning their crop development in time and space, and to identify information and data requirements. Then, we will adapt current crop simulation tools developed by CSIRO using model-order reduction techniques to develop sequential data analytics to solve these multivariable spatio-temporal processes.

You will work closely with researchers from CSIRO and Data61 who have detailed working knowledge of these existing models to improve their relevance and power for agricultural decision-making.

In the second part of the project, we will develop novel methods to analyse the kinds of multi-variable data that are increasingly available in the data-rich agricultural sector. In these methods, you will utilise the information feedback from the previous stage to ground these optimal spatio-temporal management strategies in the real world.


The supervisory team spanning Queensland University of Technology, the Insitutute for Future Environments, the ARC Centre of Excellence for Mathematical and Statistical Frontiers, Data61 and CSIRO has unique and complementary expertise to develop transdisciplinary research aimed at optimising agricultural management strategies.
The tools you develop in this PhD project will be directly useful for applications in in-season management of crops, including irrigation and nutrition management. Collaboration with CSIRO offers the rare opportunity to investigate stimulating mathematical and statistical problems in a field that will directly improve Australia's economic, nutritional, and environmental future.

We expect the outcomes of the project to include:

  • procedures for model-order reduction of spatio-temporal processes
  • formulation of optimal sequential-decision problems with application to crop development and growth management
  • analysis of system properties such as reachability and observability
  • novel sequential data analytics for multivariable spatio-temporal processes
  • novel information-feedback decision strategies with incomplete information
  • development of a benchmark management strategy for a particular crop.

We expect the work to have an impact through:

  • developing insights and understanding on modelling requirements and uncertainty
  • characterisation related to sequential data analytics and optimal feedback management strategies
  • creating new benchmarks to motivate practice change in management based on digital agriculture
  • raising interest in validated dynamical-system tools for the application to crop management
  • your ability to engage with scientists from other areas and bring new tools into agriculture

Skills and experience


You must have:

  • an undergraduate degree in mathematics, statistics, computer science, data science, information technology, or similar quanitative field
  • willingness to collaborate with scientists outside of your field
  • willingness to engage with partners from non-scientific fields (e.g. farmers)


  • statistical computing skills in R, matlab, Python, C or other programming languages
  • familiarity with statistical software, bayesian methods, spatial modelling, temporal modelling, or spatial-temporal modelling.


This PhD project is jointly funded by QUT and CSIRO. See the QUT-CSIRO Digital Agriculture PhD scholarship web page for full details, including the application process.



Contact the supervisor for more information.