Decision-making tools (KDSS)
The KDSS decision-making support tool uses automated reasoning and learning techniques for recommending efficient strategies on the monitoring of rainwater deposits and management of bathing areas.
The KDSS (knowledge-driven decision-support system) tool is being implemented in the open platform on two levels:
- The local level, which enables monitoring strategies to minimise discharges from the unit system (DSU) and untreated-rainwater leakages; to reduce the presence of solid, plastic and fibre spillages; and to determine the concentration of substances that are harmful to human health in bathing water.
- The global level that determines the strategies for operational management (general considerations with information from the global network) and bathing (such as determining bathing flags).
Local intelligence is being implemented as simple rules (by considering flows, solids in suspension, concentrations etc.) and anomaly detections, set up remotely in SCADA systems. Global intelligence is based on multi-goal optimisation for operational recommendations by considering environmental, economic and social impact as a function of costs (minimisation function).
Bathing-area recommendations are determined through a combination of deep learning (neural networks) with other learning and statistical tools (X-MEANS for achieving seasonal patterns, data-cleaning techniques such as PCA). Global intelligence also enjoys the support of Quantitative Microbial Risk Assessment (QMRA) and dispersion models. In addition, the information is captured through a semantic sensor concentrator.
Bathing-area recommendations are determined through a combination of deep learning (neural networks) with other learning and statistical tools (X-MEANS for achieving seasonal patterns, data-cleaning techniques such as PCA). Global intelligence also enjoys the support of Quantitative Microbial Risk Assessment (QMRA) and dispersion models. In addition, the information is captured through a semantic sensor concentrator.
