Abstract
The sparse nonlinear programming (SNP) problem has wide applications in signal and image processing, machine learning and finance, etc. However, the computational challenge posed by SNP has not yet been well resolved due to the nonconvex and discontinuous \(\ell _0\)-norm involved. In this paper, we resolve this numerical challenge by developing a fast Newton-type algorithm. As a theoretical cornerstone, we establish a first-order optimality condition for SNP based on the concept of strong \(\beta \)-Lagrangian stationarity via the Lagrangian function, and reformulate it as a system of nonlinear equations called the Lagrangian equations. The nonsingularity of the corresponding Jacobian is discussed, based on which the Lagrange–Newton algorithm (LNA) is then proposed. Under mild conditions, we establish the locally quadratic convergence and its iterative complexity estimation. To further demonstrate the efficiency and superiority of our proposed algorithm, we apply LNA to two specific problems arising from compressed sensing and sparse high-order portfolio selection, in which significant benefits accrue from the restricted Newton step.
Similar content being viewed by others
Notes
HTP is available at:https://github.com/foucart/HTP.
NIHT, GP and OMP are available at https://www.southampton.ac.uk/engineering/about/staff/tb1m08.page#software. We use the version sparsity_0_5 in which NIHT, GP and OMP are called hard_l0_Mterm, greed_gp and greed_omp.
CoSaMP and SP are available at:http://media.aau.dk/null_space_pursuits/2011/07/a-few-corrections-to-cosamp-and-sp-matlab.html.
References
Beck, A., Eldar, Y.C.: Sparsity constrained nonlinear optimization: optimality conditions and algorithms. SIAM J. Optim. 23(3), 1480–1509 (2013)
Beck, A., Hallak, N.: On the minimization over sparse symmetric sets: projections, optimality conditions, and algorithms. Math. Oper. Res. 41(1), 196–223 (2015)
Beck, A., Vaisbourd, Y.: The sparse principal component analysis problem: Optimality conditions and algorithms. J. Optim. Theory Appl. 170(1), 119–143 (2016)
Blumensath, T., Davies, M.E.: Gradient pursuits. IEEE Trans. Signal Process. 56(6), 2370–2382 (2008)
Blumensath, T., Davies, M.E.: Iterative hard thresholding for compressed sensing. Appl. Comput. Harmon. Anal. 27(3), 265–274 (2009)
Blumensath, T., Davies, M.E.: Normalized iterative hard thresholding: guaranteed stability and performance. IEEE J. Sel. Top. Signal Process. 4(2), 298–309 (2010)
Boudt, K., Lu, W., Peeters, B.: Higher order comoments of multifactor models and asset allocation. Financ. Res. Lett. 13, 225–233 (2015)
Červinka, M., Kanzow, C., Schwartz, A.: Constraint qualifications and optimality conditions for optimization problems with cardinality constraints. Math. Program. 160(1), 353–377 (2016)
Chen, J., Gu, Q.: Fast Newton hard thresholding pursuit for sparsity constrained nonconvex optimization. In: Proceedings of the 23rd ACM SIGKDD international conference on knowledge discovery and data mining, pp. 757–766 (2017)
Chen, X., Ge, D., Wang, Z., Ye, Y.: Complexity of unconstrained \(l_2-l_p\) minimization. Math. Program. 143(1–2), 371–383 (2014)
Dai, W., Milenkovic, O.: Subspace pursuit for compressive sensing signal reconstruction. IEEE Trans. Inf. Theory 55(5), 2230–2249 (2009)
Donoho, D.L.: Compressed sensing. IEEE Trans. Inf. Theory 52(4), 1289–1306 (2006)
Elad, M.: Sparse and Redundant Representations. Springer, New York (2010)
Elad, M., Figueiredo, M.A., Ma, Y.: On the role of sparse and redundant representations in image processing. Proc. IEEE 98(6), 972–982 (2010)
Fan, J., Li, R.: Variable selection via nonconcave penalized likelihood and its Oracle properties. J. Am. Stat. Assoc. 96(456), 1348–1360 (2001)
Foucart, S.: Hard thresholding pursuit: an algorithm for compressive sensing. SIAM J. Numer. Anal. 49(6), 2543–2563 (2011)
Gao, J., Li, D.: Optimal cardinality constrained portfolio selection. Oper. Res. 61(3), 745–761 (2013)
Gotoh, J.Y., Takeda, A., Tono, K.: DC formulations and algorithms for sparse optimization problems. Math. Program 169(1), 141–176 (2018)
Han, S.P.: Superlinearly convergent variable metric algorithms for general nonlinear programming problems. Math. Program. 11, 263–282 (1976)
Koh, K., Kim, S.J., Boyd, S.: An interior-point method for large-scale \(\ell _1\)-regularized logistic regression. J. Mach. Learn. Res. 8, 1519–1555 (2007)
Kyrillidis, A., Becker, S., Cevher, V., Koch, C.: Sparse projections onto the simplex. In: Proceedings of the 30th international conference on machine learning, Atlanta, Georgia, USA, 2013, vol. 28, pp. 235–243 (2013)
Lu, Z.: Optimization over sparse symmetric sets via a nonmonotone projected gradient method. arXiv:1509.08581 (2015)
Lu, Z., Zhang, Y.: Sparse approximation via penalty decomposition methods. SIAM J. Optim. 23(4), 2448–2478 (2013)
Luo, Z., Sun, D., Toh, K.C., Xiu, N.: Solving the OSCAR and SLOPE models using a semismooth Newton-based augmented Lagrangian method. J. Mach. Learn. Res. 20(106), 1–25 (2019)
Luo, Z.Q., Pang, J.S., Ralph, D. Piecewise sequential quadratic programming for mathematical programs with nonlinear complementarity constraints. In: Migdalas, A., Pardalos, P.M., Värbrand, P. (eds.) Multilevel Optimization: Algorithms and Applications. Nonconvex Optimization and Its Applications, vol 20. Springer, Boston, MA (1998)
Misra, J.: Interactive exploration of microarray gene expression patterns in a reduced dimensional space. Genome Res. 12(7), 1112–1120 (2002)
Natarajan, B.K.: Sparse approximate solutions to linear systems. SIAM J. Comput. 24(2), 227–234 (1995)
Needell, D., Tropp, J.A.: CoSaMP: iterative signal recovery from incomplete and inaccurate samples. Appl. Comput. Harmon. Anal. 26(3), 301–321 (2009)
Negahban, S.N., Ravikumar, P., Wainwright, M.J., Yu, B., et al.: A unified framework for high-dimensional analysis of m-estimators with decomposable regularizers. Stat. Sci. 27(4), 538–557 (2012)
Pan, L., Luo, Z., Xiu, N.: Restricted Robinson constraint qualification and optimality for cardinality-constrained cone programming. J. Optim. Theory Appl. 175(1), 104–118 (2017)
Pan, L., Xiu, N., Fan, J.: Optimality conditions for sparse nonlinear programming. Sci. China Math. 60(5), 759–776 (2017)
Pan, L., Xiu, N., Zhou, S.: On solutions of sparsity constrained optimization. J. Oper. Res. Soc. China 3(4), 421–439 (2015)
Pan, L., Zhou, S., Xiu, N., Qi, H.D.: Convergent iterative hard thresholding for sparsity and nonnegativity constrained optimization. Pacif. J. Optim. 13(2), 325–353 (2017)
Pati, Y.C., Rezaiifar, R., Krishnaprasad, P.S.: Orthogonal matching pursuit: recursive function approximation with applications to wavelet decomposition. In: Proceedings of 27th Asilomar Conference on Signals, Systems and Computers, pp. 40–44. IEEE (1993)
Powell, M.J.D.: Restart procedures for the conjugate gradient method. Math. Program. 12, 241–254 (1977)
Tropp, J.A., Gilbert, A.C.: Signal recovery from random measurements via orthogonal matching pursuit. IEEE Trans. Inf. Theory 53(12), 4655–4666 (2007)
Wang, J., Deng, Z., Zheng, T., So, A.M.C.: Sparse high-order portfolios via proximal DCA and SCA. In: ICASSP 2021–2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 5425–5429 (2021)
Wilson, R.B.: A simplicial algorithm for concave programming. Ph.D. thesis, Graduate School of Business Administration, Harvard University (1963)
Xu, F., Lu, Z., Xu, Z.: An efficient optimization approach for a cardinality-constrained index tracking problem. Optim. Methods Softw. 31, 258–271 (2016)
Yin, P., Lou, Y., He, Q., Xin, J.: Minimization of \(\ell _{1-2}\) for compressed sensing. SIAM J. Sci. Comput. 37(1), 536–563 (2015)
Yuan, X., Li, P., Zhang, T.: Gradient hard thresholding pursuit. J. Mach. Learn. Res. 18(166), 1–43 (2018)
Yuan, X., Liu, Q.: Newton greedy pursuit: a quadratic approximation method for sparsity-constrained optimization. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, pp. 4122–4129 (2014)
Yuan, X., Liu, Q.: Newton-type greedy selection methods for \(\ell _0\)-constrained minimization. IEEE Trans. Pattern Anal. Mach. Intell. 39(12), 2437–2450 (2017)
Zhou, S., Xiu, N., Qi, H.: Global and quadratic convergence of Newton hard-thresholding pursuit. J. Mach. Learn. Res. 22, 1–45 (2021)
Zhou, S., Xiu, N., Wang, Y., Kong, L., Qi, H.D.: A null-space-based weighted \(\ell _1\) minimization approach to compressed sensing. Inf. Inference J. IMA 5(1), 76–102 (2016)
Zhou, T., Tao, D., Wu, X.: Manifold elastic net: a unified framework for sparse dimension reduction. Data Min. Knowl. Disc. 22(3), 340–371 (2011)
Zou, H., Hastie, T., Tibshirani, R.: Sparse principal component analysis. J. Comput. Graph. Stat. 15(2), 265–286 (2006)
Acknowledgements
We would like to thank AE and the two referees for their valuable comments, which have helped to shorten the paper and greatly improved its presentation. We also thank Dr. Shenglong Zhou of Imperial College, London for his great support on the numerical experiments.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This research was partially supported by the National Natural Science Foundation of China (11771038, 11971052, 12011530155) and Beijing Natural Science Foundation (Z190002)
Rights and permissions
About this article
Cite this article
Zhao, C., Xiu, N., Qi, H. et al. A Lagrange–Newton algorithm for sparse nonlinear programming. Math. Program. 195, 903–928 (2022). https://doi.org/10.1007/s10107-021-01719-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10107-021-01719-x
Keywords
- Sparse nonlinear programming
- Lagrangian equation
- The Newton method
- Locally quadratic convergence
- Application