Using multiobjective optimization to reconstruct interferometric data (II): polarimetry and time dynamics

Alejandro Mus,Hendrik Müller,Ivan Martí-Vidal,Andrei Lobanov

2024-01-24

Abstract:In Very Long Baseline Interferometry (VLBI), signals from multiple antennas combine to create a sparsely sampled virtual aperture, its effective diameter determined by the largest antenna separation. The inherent sparsity makes VLBI imaging an ill-posed inverse problem, prompting the use of algorithms like the Multiobjective Evolutionary Algorithm by Decomposition (MOEA/D), as proposed in the first paper of this series. This study focuses on extending MOEA/D to polarimetric and time dynamic reconstructions, particularly relevant for the VLBI community and the Event Horizon Telescope Collaboration (EHTC). MOEA/D's success in providing a unique, fast, and largely unsupervised representation of image structure serves as the basis for exploring these extensions. The extension involves incorporating penalty terms specific to total intensity imaging, time-variable, and polarimetric variants within MOEA/D's multiobjective, evolutionary framework. The Pareto front, representing non-dominated solutions, is computed, revealing clusters of proximities. Testing MOEA/D with synthetic datasets representative of EHTC's main targets demonstrates successful recovery of polarimetric and time-dynamic signatures despite sparsity and realistic data corruptions. MOEA/D's extension proves effective in the anticipated EHTC setting, offering an alternative and independent claim to existing methods. It not only explores the problem globally but also eliminates the need for parameter surveys, distinguishing it from Regularized Maximum Likelihood (RML) methods. MOEA/D emerges as a novel and useful tool for robustly characterizing polarimetric and dynamic signatures in VLBI datasets with minimal user-based choices. Future work aims to address the last remaining limitation of MOEA/D, specifically regarding the number of pixels and numerical performance, to establish it within the VLBI data reduction pipeline.

Instrumentation and Methods for Astrophysics,Astrophysics of Galaxies,Optimization and Control

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: in very long baseline interferometry (VLBI), due to the sparsity and complexity of observational data, traditional imaging methods are difficult to effectively reconstruct polarization and time - dynamic characteristics. The author proposes to extend the multi - objective evolutionary algorithm (MOEA/D) to achieve efficient reconstruction of polarization and time - dynamic characteristics, especially for the needs of the Event Horizon Telescope Collaboration (EHTC).

Specifically, the paper aims to:

1. **Deal with the inverse problem in VLBI imaging**: Due to the sparsity of VLBI data and noise interference, imaging is an ill - posed inverse problem. Traditional methods such as CLEAN and its variants rely on human intervention and parameter tuning, while MOEA/D provides a fast, unsupervised method.

2. **Extend MOEA/D to polarization and time - dynamic reconstruction**: By introducing specific penalty terms, MOEA/D is applied to total intensity imaging, time - varying and polarization variants. This includes dealing with calibration problems such as polarization leakage and feed rotation, and revealing non - dominated solution clusters through the Pareto front.

3. **Improve the imaging ability of EHTC**: Provide a new tool for EHTC that can robustly characterize polarization and dynamic features in VLBI datasets, reducing the impact of user selection.

4. **Solve the computational complexity of high - dimensional problems**: Currently, MOEA/D faces a performance bottleneck when dealing with high - dimensional problems, and the author is developing new techniques to overcome this limitation, such as particle swarm optimization - based methods.

In summary, the goal of the paper is to provide the VLBI community with a more powerful and automated way to reconstruct polarization and time - dynamic images by improving and extending the MOEA/D algorithm, especially in the face of challenges in millimeter - wave global VLBI observations.