Project Results


EVOSEC project aimed at applying and extending new generations of heuristics and meta-heuristics for optimization & search issues in security and reliability.

The project results include:

  1. The study and development of the Artificial Immune System paradigm applied to Intrusion Detection. The approach consisted in clustering TCP-IP network traffic by the means of Artificial Recognition balls and then classifying it (regular traffic vs. attack) using Gene Expression Programming. The framework has been validated through the use of generated traffic simulated under NS-2. An open source code has been produced for the gene expression library and is available here: https://gforge.uni.lu/projects/libgep/
  2. The study and development of the notion of Trust Management in ad hoc networks. A game-theoretical approach has been designed for modeling trust management on ad hoc networks. The local strategies are tuned using evolutionary approach and validated on various case studies.
    Additional to that spanning trees and Dominant Sequences have been used as a base construct for providing reliable management structure on ad-hoc networks.
  3. The study and development of a new generation of heuristics for reliable scheduling on p2p environments. A new generation of algorithms mixing exact approaches (B&B) and approximated ones (Genetic Algorithms) has been designed. This approach demonstrated its scalability through the use of Grid 5000 and succeeded to break existing records for the Q3AP problem.