DAYDREAMS will address these goals through the design, development, and integration of three technological pillars:
- AI and Machine Learning for asset management prescriptions based on asset status nowcasting and forecasting. These technologies will be targeted to model the entire maintenance process, through the use of both endogenous and exogenous (e.g., environmental) data, including asset-related physical models of the phenomena and human behaviour/decisions/actions, which holistically describe or affect the asset management process;
- Multi-objective Optimisation (including AI- and stochastic-based methods). These technologies will be targeted to prescribe optimal decisions to railway stakeholders, by ranking a list of possible options, together with related risks and uncertainties, taking into account both stakeholders’ and maintenance process metrics and constraints, taking into account stakeholders’ KPIs, preferences, and constraints.
- Context-driven HMI. These technologies will allow improving the effectiveness of information transfer (e.g. prescribed action with associated KPIs and uncertainties) to decision-makers and will allow the collection of stakeholders’ behaviour, as to obtain an effective risk-aware human-in-the-loop integrated system.
In order to prove the effectiveness and the added business value of the proposed approach, DAYDREAMS will target the development of an IAMS prototype, integrating the three DAYDREAMS pillars, and validating it by means of scenarios and data related to the railway asset management process.