Simulation of Drinking Water Infrastructures Through Artificial Intelligence-Based Modelling for Sustainability Improvement

The development of control systems for critical infrastructures requires testing and validating the proposals before using them in real environments. This work proposes the development of a new control system with an approach based on sustainability, which uses multi-agent systems as a basis, and which breaks away from traditional proposals focused on optimising energy costs. This new approach requires a thorough validation before its possible deployment, as it is based on distributed components that make independent decisions to generate complex emergent behaviour. In order to test its viability, a simulator has also been developed alongside the control system, which allows the behaviour of each agent to be analysed by subjecting it to tests using real data from the scenario to be controlled. Through this tool it is possible to observe each agent in the fulfilment of its functions, validate its behaviour, and check that the control system guarantees the supply of drinking water to a city, using the data obtained from that city as input. Through the simulator it is possible to analyse and represent different configurations of the control system over an infrastructure, thus being able to select the best option for the environment.