Tailoring a Genetic Algorithm for searching Shortest Lattice Vector in NTRU Lattices

The NTRU cryptosystem relies on the hardness of the Shortest Vector Problem (SVP), a central challenge in lattice-based cryptography. While genetic algorithms (GAs) have shown promise in solving SVP on generic lattices, little attention has been given to applying it to cryptographic structures like NTRU. In this work, we investigate the suitability of a GA specifically tailored to NTRU lattices. Our approach involves searching outside the lattice and analyzing a cost function whose behavior appears smoother when using BKZ-reduced bases. We present initial observations and outline directions for future research.