Delays in acting in collapsing ecosystems can be catastrophic. With every passing year, the chances that the ecosystem has progressed past some point of no return increases. Yet the research and development needed to develop a new technology can take a long time. Balance between these two dynamic processes is needed to determine the optimal length and effort for developing new technologies. This project will develop a method for finding the optimal schedule for developing technological readiness, social acceptability, a diminishing ecological state, and a closing window of effectiveness.
In this project, the student will explore the relationship between the speed of potential ecosystem collapse and the optimal readiness of the technology. If the window in which the technology will be effective is long, the optimal strategy is likely to develop the technology for longer; the opposite is likely if the window is short. The shape of that trade-off will inform timelines for intermediate cases.
Optimal R&D schedules for new conservation actions are a state-based optimization problem. Based on the state of the environment and the state of the technology, decision-makers must choose each year whether to continue developing, or deploy the action. The student will construct a Markov decision process mapping the state of the technology and the environment to an optimal strategy for each year – either continue R&D (technical, social, or both) or deploy. They will then use Stochastic Dynamic Programming and backwards iteration to find the optimal schedule for a case study technology for managing the Great Barrier Reef.
The explicit aims of the project will depend on the level and background of the student. This project forms part of larger research projects focused on the Great Barrier Reef and Antarctica with other government and university collaborators and stakeholders, and could be used as a launching pad for further research with these groups.
Skills and experience
- Advanced quantitative skills, or a demonstrated interest in developing them
- programming skills (any language), or a demonstrated interest in developing them
- mathematics students with backgrounds in Statistics, Operations Research, Applied Computational Maths, decision Science are particularly encouraged, as are students with a background in economics or computer science. However, this is not completely necessary
- good interpersonal and written communication skills are necessary, as well as a willingness to meet with and collaborate with scientists outside of maths.
- decision science
- optimal scheduling
- ecosystem management
- biodiversity conservation
- markov decision processes
Contact the supervisor for more information.