I started my research in ENEA working on the application of evolutionary algorithms and neural networks to modeling and optimization problems. Currently, I’m working on Climate Data Analysis and Climate Services. My main research topics are the following.
- Energy Production/Demand Forecasting with Statistical Methods
- Seasonal Forecast Application for Energy Systems
- Other Research Topics about NNs and EC
Energy Production/Demand Forecasting with Statistical Methods
Electricity demand forecasting is gaining a lot of attention due to the raising interest for Smart Grids, or more in general Intelligent Grids, in fact the necessity to make the ratio cost/efficiency as lower as possible leads to actions aimed to shape the energy peak demands (called peak shaping or load shifting). On larger geographical scales, the interest for an accurate load forecast is growing due to the deregulation of electricity markets and the spread of renewable energy sources that make the scheduling and dispatching more difficult. In fact an accurate prediction gives the possibility to plan the energy production or to manage effectively supplies (energy storage).
Furthermore, the diffusion of Renewable Energies has raised the necessity of accurate local information about the actual production and its forecast for scheduling purposes. This activity is mainly focused on solar- and hydro-power.
This research may be performed with a wide variety of methods: statistical methods (Box-Jenkins, Holt-Winters etc), neural networks, support vector machines, data mining approaches.
Open projects and collaborations:
- Short-term Electricity Demand Forecasting on Italian National Grid [with TERNA and UTMEA-CLIM]
- Solar-power modelling and forecast [with TERNA and UTMEA-CLIM]
Seasonal Forecast Application for Energy Systems
Climate seasonal forecasts give the possibility to predict weather parameters anomalies with a lead time of several months. Given the strong link between weather and energy systems, I’ve been working on the use of information provided by seasonal forecasts for:
- Electricity Demand forecast with a particular attention to extreme weather events (cold-spells and heat-waves)
- Renewable energies production, in particular hydro- and solar-power
Open projects and collaborations:
- Seasonal Electricity Demand Forecasting on Italian National Grid [with TERNA and UTMEA-CLIM]
- FP7 EU-PROJECT EUPORIAS and SPECS
- Hydro-power production on seasonal to decadal time-scales [with Uni Politecnica of Bucarest]
Other Research Topics about NNs and EC
I’ve been working with Neural Networks since the beginning of my Ph.D. and I’ve developed a genuine interest for the optimization of Neural Networks structure and topology. Currently, I’m collaborating with the University of Milan for an investigation about the application of evolutionary computation on neural network design.
One of the best features of evolutionary algorithms (like Genetic Algorithms) is the possibility of a parallel and distributed implementation, for these reason during the last decades several algorithm architectures have been developed.I’ve been working, in collaboration with the University of Zaragoza and the University of Roma Tre, on the analogies between Spatially-Structured Evolutionary Algorithms (SSGAs), like Cellular Genetic Algorithms (CGA), and the epidemiology models used for epidemic spreading.