Algorithms for finite groups: quasipolynomial normalisers through group structure

P vs NP, one of the seven Millennium Mathematics Problems, drives fundamental questions in complexity theory. The graph isomorphism problem (GI) and related challenges in computational group theory occupy a fascinating position in this landscape—potentially neither in P nor NP-complete—making them fascinating test cases for understanding computational hardness. Babai's breakthrough quasipolynomial algorithm for GI uses advanced group theory, demonstrating how group-theoretic approaches can revolutionise computational complexity.This project targets the NORM problem: given subgroups G and H of a fixed symmetric group, construct the normaliser of G in H. Both GI and NORM are related to a family of polynomially equivalent group theory problems, the Luks class. GI reduces to this class, and all problems in the Luks class reduce to NORM, positioning it at the apex of the hierarchy. I will prove sharper complexity bounds and develop efficient practical algorithms using an innovative approach that utilises deep structural properties of finite groups.Current general bounds for NORM show simply exponential complexity. Recent advances have achieved quasipolynomial-time solutions for primitive groups, but vast gaps remain for general permutation groups. I will fill these gaps by combining, in an innovative way, my expertise in generation and crown theory with my supervisor's computational algebra expertise. I will develop polynomial algorithms for primitive groups (WP1) and quasipolynomial ones for the general case (WP2), and release optimised algorithms in GAP/MAGMA (WP3).AlFiGs will innovately apply structural group theory to computational problems, while building new algebraic tools, and delivering immediate benefits to researchers through open-source implementations that accelerate progress in computational group theory and in its endless applications.This will make me a competitive independent researcher, who can aspire to become an expert in the field.