Motivation. Bushfires can cause poor air quality, which can affect human and animal health, and can have long-lasting impacts on soil and water quality. Bushfires can also have devastating impacts on plants, animals and ecosystems. Furthermore, the changes in climate (caused by the drastic increase in carbon dioxide level in the atmosphere) create warmer, drier conditions which are boosting the frequency and intensity of the bushfires. Consequently, development of methods to monitor and control bushfires is very essential.
Computational Perspective. There has been an ongoing effort in the field of computer science to design effective methods to monitor and control the spread of wildfires, using computer vision tools, wireless sensor networks, unmanned aerial vehicles, and machine learning algorithms. This project aims to leverage the graph theory and algorithm design techniques to make advancement in our knowledge on the control of bushfires.
Bushfire Spread Model. Consider an n*n grid, where initially each cell is unburned, burning, or burned. In each discrete-time round: (i) a burned cell remains burned, (ii) a burning cell switches to burned if it has been in the burning state during the last t rounds, (iii) an unburned cell becomes burning with some probability p if it is adjacent to a burning cell. The values of t and p can be set according to the set-up that model attempts to mimic. One can, of course, add different parameters to the aforementioned basic model and consider other network structures to capture the real-world bushfires spread process more comprehensively.
Please feel free to contact me if you have any questions. I would be glad to have a discussion on whether this project suits you.