Exact Sparse Approximation Problems via Mixed-Integer Programming: Formulations and Computational Performance - Archive ouverte HAL Access content directly
Journal Articles IEEE Transactions on Signal Processing Year : 2016

Exact Sparse Approximation Problems via Mixed-Integer Programming: Formulations and Computational Performance

Abstract

Sparse approximation addresses the problem of approximately fitting a linear model with a solution having as few non-zero components as possible. While most sparse estimation algorithms rely on suboptimal formulations, this work studies the performance of exact optimization of problems through Mixed-Integer Programs (MIPs). Nine different sparse optimization problems are formulated based on or data misfit measures, and involving whether constrained or penalized formulations. For each problem, MIP reformulations allow exact optimization, with optimality proof, for moderate-size yet difficult sparse estimation problems. Algorithmic efficiency of all formulations is evaluated on sparse deconvolution problems. This study promotes error-constrained minimization of the norm as the most efficient choice when associated with and misfits, while the misfit is more efficiently optimized with sparsity-constrained and sparsity-penalized problems. Then, exact optimization is shown to outperform classical methods in terms of solution quality, both for over-and under-determined problems. Finally, numerical simulations emphasize the relevance of the different fitting possibilities as a function of the noise statistical distribution. Such exact approaches are shown to be an efficient alternative, in moderate dimension, to classical (suboptimal) sparse approximation algorithms with data misfit. They also provide an algorithmic solution to less common sparse optimization problems based on and misfits. For each formulation, simulated test problems are proposed where optima have been successfully computed. Data and optimal solutions are made available as potential benchmarks for evaluating other sparse approximation methods.
Fichier principal
Vignette du fichier
MIP_sparse.pdf (662.51 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01254856 , version 1 (12-01-2016)

Identifiers

Cite

Sébastien Bourguignon, Jordan Ninin, Hervé Carfantan, Marcel Mongeau. Exact Sparse Approximation Problems via Mixed-Integer Programming: Formulations and Computational Performance. IEEE Transactions on Signal Processing, 2016, 64 (6), pp.1405-1419. ⟨10.1109/TSP.2015.2496367⟩. ⟨hal-01254856⟩
805 View
1661 Download

Altmetric

Share

Gmail Facebook X LinkedIn More