The research focus of this project is urban water demand management with the objective of understanding how to capture the long- and short-term demand changes and their responses to external stressor under heterogeneous geographical contexts, and relationships with determinants, exploring whether we could accurately forecast future water demands also under uncertain scenarios, and assessing how real short- to medium-term demand forecasting influence control of water supply network.
Methods applied
Quantitative longitudinal performance data
Quantitative data analysis
Analytical tools & literature
Machine learning/deep learning models
Control optimization tool
Reflections and recommendations by the researcher
Machine learning is not a panacea, the black-box nature of the model needs to be solved with new structure and better understanding of real natural processes.