Welcome to Matteo De Felice's Homepage
Research works
These are my current research works:
Fitness Approximation for Computationally Expensive Problems
In collaboration with ENEA. Real-world problems need software simulators that often are really expensive both computionally and economically. Black-box optimization methodologies (like Evolutionary Algorithms, Swarm-based algorithms and so on) can use an approximated fitness function to drastically reduce the number of calls of the performance functions and therefore of the simulator.
Soft Computing and Immuno-based algorithm for Anomaly Detection
In collaboration with DIA and University of Milan. Different methodologies are used to detect anomalies in different systems (e.g. Internet): ANNs, PSO, Multi-Objective Evolutionary Algorithms, AIS.
Spatially-Structured Evolutionary Algorithms
In collaboration with DIA. Studying graph-based EAs and the influence of topology on diversity and convergence speed
