The Gross-Pitaevskii equation and eigenvector nonlinearities: numerical methods and algorithms (Q6972399)

This work is a substantial complete review of the state-of-the-art of numerical methods for computing (ground) eigenpairs of the Gross-Pitaevskii eigenvalue problem (GPEVP). The scope of the work involves the case of bounded convex domains of 1, 2 and 3 dimensions, where the potential is positive and bounded. The work starts with an overview of the formulation of the GPEVP, its context in the study of Bose-Einstein condensates, formulation of the optimisation problem and its underlying nonlinear partial differential equation and discretisation via finite elements. Then, in the main body of the review, the authors give a systematic treatment of the methods available, under two overarching headings: algebraic viewpoint and variational viewpoint.\N\NUnder the algebraic approach heading, the authors survey the variety of numerical linear algebra techniques surrounding a Newton's method formulation for the ground eigenpair. In it, the authors describe both general and GPEVP-specific frameworks, including the self-consistent field iteration, inverse iteration implicit iterative, sequential quadratic and other methods.\N\NUnder the variational viewpoint heading, they describe GPEVP-specific continuous and discrete gradient flow approaches, Riemannian conjugate gradient methods, and a final description of what is currently known about pre-conditioning and damping in this variational context.\N\NThe presentation is illustrated through the text, by benchmark convergence comparisons on the main approaches for a finite element discretisation of the PDE, posed on a 2D rectangle with a specific trapping potential. A proper outlook of the theory and a complete list of references are included in the last part of the work.