[Paper Review] A fast direct method of mass reconstruction for gravitational lenses

M. Lombardi, G. Bertin|arXiv (Cornell University)|Jun 7, 1999

Astrophysical Phenomena and Observations1 references3 citations

TL;DR

This paper presents a fast direct method for reconstructing mass distributions in gravitational lensing by solving a variational principle efficiently, reducing computation time by 100 to 1000 times compared to existing methods—enabling sub-second reconstruction on standard workstations for grids up to N = 400. The approach makes large-scale statistical and cosmological simulations of lensing observations feasible for the first time with practical speed.

ABSTRACT

Statistical analyses of observed galaxy distortions are often used to reconstruct the mass distribution of an intervening cluster responsible for gravitational lensing. In current projects, distortions of thousands of source galaxies have to be handled efficiently; much larger data bases and more massive investigations are envisaged for new major observational initiatives. In this article we present an efficient mass reconstruction procedure, a direct method that solves a variational principle noted in an earlier paper, which, for rectangular fields, turns out to reduce the relevant execution time by a factor from 100 to 1000 with respect to the fastest methods currently used, so that for grid numbers N = 400 the required CPU time on a good workstation can be kept within the order of 1 second. The acquired speed also opens the way to some long-term projects based on simulated observations (addressing statistical or cosmological questions) that would be, at present, practically not viable for intrinsically slow reconstruction methods.

Motivation & Objective

• To address the computational bottleneck in reconstructing mass distributions from galaxy shear data in gravitational lensing.
• To enable efficient processing of large datasets involving thousands of source galaxies, as required by upcoming observational projects.
• To develop a method that drastically reduces computation time while maintaining accuracy for mass reconstruction in lensing studies.
• To make long-term simulations of lensing observations—previously infeasible due to slow reconstruction—practically viable.

Proposed method

• The method solves a variational principle previously noted in the literature, reformulating the mass reconstruction as a constrained minimization problem.
• For rectangular fields, the method exploits a fast Fourier transform (FFT)-based algorithm to accelerate the solution of the underlying linear system.
• The approach avoids iterative solvers by directly computing the solution through a closed-form expression derived from the variational principle.
• The algorithm is optimized for regular grids, allowing efficient matrix-vector operations using FFTs.
• The method is implemented on standard workstations and achieves sub-second performance for N = 400 grids.

Experimental results

Research questions

• RQ1 Can a direct method be developed to reconstruct mass distributions in gravitational lensing faster than current iterative approaches?
• RQ2 What is the maximum speedup achievable in mass reconstruction using a direct solution to the variational principle?
• RQ3 Can the method scale to large datasets with thousands of source galaxies and still maintain real-time performance?
• RQ4 Is it now feasible to perform long-term simulations of lensing observations using this reconstruction speed?

Key findings

• The method reduces CPU time by a factor of 100 to 1000 compared to the fastest existing methods for mass reconstruction.
• For grid sizes up to N = 400, the required computation time is kept within approximately 1 second on a good workstation.
• The direct solution via the variational principle enables practical use of simulated lensing observations for statistical and cosmological investigations.
• The method's efficiency is achieved through FFT-based acceleration, making it suitable for large-scale data processing.