Abstract
Recently, motivated by works arising from large-scale linear systems, some stochastic and deterministic nonlinear Kaczmarz (NK) methods have been developed to solve large-scale nonlinear system of equations. In this paper, based on row sample with an approximate maximum residual control criterion, we propose a pseudoinverse-free block maximum residual nonlinear Kaczmarz (FBMRNK) method for solving large-scale nonlinear system of equations. Then we show that FBMRNK is a variant of the sketched Newton–Raphson (SNR) method with an adaptive sketching matrix. Furthermore, using this connection we establish the global convergence theory of FBMRNK with \(\mu\)-strongly quasi-convexity or star-convexity. Finally, numerical results are provided to demonstrate superior performance of FBMRNK to the stochastic or deterministic nonlinear Kaczmarz-type methods.
Similar content being viewed by others
Data availability
All data generated or analysed during this study are included in this article.
References
Cands, E.J., Li, X., Soltanolkotabi, M.: Phase retrieval via Wirtinger flow: theory and algorithms. IEEE Trans. Inf. Theory. 61, 1985–2007 (2015)
Ortega, J.M., Rheinboldt, W.C.: Iterative solution of nonlinear equations in several variables. In: Classics Appl. Math., vol. 30. SIAM, Philadelphia (2000)
Torres, G.L., Quintana, V.H.: Optimal power flow by a nonlinear complementarity method. IEEE Trans. Power Syst. 15, 1028–1033 (2000)
Boyd, S., Boyd, S.P., Vandenberghe, L.: Convex Optimization. Cambridge University Press, Cambridge (2004)
Wright, S., Nocedal, J.: Numerical Optimization, 2nd edn. Springer Science and Business Media, Berlin (2006)
Kawaguchi, K.: Deep learning without poor local minima. Adv. Neural Inf. Process. Syst. 29, 586–594 (2016)
Brewster, M.E., Kannan, R.: Nonlinear successive over-relaxation. Numer. Math. 44, 309–315 (1984)
Brewster, M.E., Kannan, R.: A computational process for choosing the relaxation parameter in nonlinear SOR. Computing 37, 19–29 (1986)
Ecker, A., Gross, D.: A system of simultaneous non-linear equations in three-thousand variables. J. Comput. Phys. 64, 246–252 (1986)
Agarwal, N., Bullins, B., Hazan, E.: Second-order stochastic optimization for machine learning in linear time. J. Mach Learn Res. 18, 1–40 (2017)
Bollapragada, R., Byrd, R.H., Nocedal, J.: Exact and inexact subsampled Newton methods for optimization. IMA J. Numer Anal. 39, 545–578 (2018)
Gower, R., Koralev, D., Lieder, F., Richtárik, P.: RSN: randomized subspace Newton. arXiv preprint arXiv:1905.10874 (2019)
Khorasani, F.R., Mahoney, M.W.: Sub-sampled Newton methods I: globally convergent algorithms. arXiv preprint arXiv:1601.04737 (2016)
Kovalev, D., Mishchenko, K., Richtárik, P.: Stochastic Newton and cubic Newton methods with simple local linear-quadratic rates. arXiv preprint arxiv:1912.01597 (2019)
Dai, Y.-H.: Convergence properties of the BFGS algorithm. SIAM J. Optim. 13, 693–701 (2002)
Byrd, R.H., Hansen, S.L., Nocedal, J., Singer, Y.: A stochastic quasi-Newton method for large-scale optimization. SIAM J. Optim. 26, 1008–1031 (2016)
Schraudolph, N.N., Yu, J., Gunter, S.: A stochastic quasi-Newton method for online convex optimization. In: International Conference on Artificial Intelligence and Statistics, Society for Artificial Intelligence and Statistics. pp 436–443 (2007)
Kantorovitch, L.: The method of successive approximation for functional equations. Acta Math. 71, 63–97 (1939)
Deuflhard, P.: Newton methods for nonlinear problems: affine invariance and adaptive algorithms. In: Springer Series in Computational Mathematics, vol. 35. Springer, Berlin (2004)
Yuan, R., Lazaric, A., Gower, R.M.: Sketched Newton–Raphson. SIAM J. Optim. 32, 1555–1583 (2022)
Gower, R.M., Richtárik, P.: Randomized iterative methods for linear systems. SIAM J. Matrix Anal. Appl. 36, 1660–1690 (2015)
Wang, Q.-F., Li, W.-G., Bao, W.-D.: Nonlinear Kaczmarz algorithms and their convergence. J. Comput. Appl. Math. 399, 113720 (2022)
Jin, B.T., Zhou, Z.-H., Zou, J.: On the convergence of stochastic gradient descent for nonlinear ill-posed problems. SIAM J. Optim. 30, 1421–1450 (2020)
Kaczmarz, S.: Angenaherte auflosung von systemen linearer glei-chungen. Bull. Intern. Acad. Pol. Sic. Lett., Cl. Sci. Math. Nat. 355–357 (1937)
Strohmer, T., Vershynin, R.: A randomized Kaczmarz algorithm with exponential convergence. J. Fourier Anal. Appl. 15, 262–278 (2009)
Eldar, Y.C., Needell, D.: Acceleration of randomized Kaczmarz method via the Johnson–Lindenstrauss lemma. Numer. Algorithms 58, 163–177 (2011)
Leventhal, D., Lewis, A.S.: Randomized methods for linear constraints: convergence rates and conditioning. Math. Oper. Res. 35, 641–654 (2010)
Liu, J., Wright, S.: An accelerated randomized Kaczmarz algorithm. Math. Comput. 85, 153–178 (2016)
Ma, A., Needell, D., Ramdas, A.: Convergence properties of the randomized extended Gauss–Seidel and Kaczmarz methods. SIAM J. Matrix Anal. Appl. 36, 1590–1604 (2015)
Needell, D.: Randomized Kaczmarz solver for noisy linear systems. BIT 50, 395–403 (2010)
Elfving, T.: Block-iterative methods for consistent and inconsistent linear equations. Numer. Math. 35, 1–12 (1980)
Eggermont, G.T., Herman, P.P.B., Lent, A.: Iterative algorithms for large partitioned linear systems. Linear Algebra Appl. 40, 37–67 (1981)
Needell, D., Tropp, J.-A.: Paved with good intentions: analysis of a randomized block Kaczmarz method. Linear Algebra Appl. 441, 199–221 (2014)
Niu, Y.-Q., Zheng, B.: A greedy block Kaczmarz algorithm for solving large-scale linear systems. Appl. Math. Lett. 104, 106294 (2020)
Moorman, J.D., Tu, T.K., Molitor, D., Needell, D.: Randomized Kaczmarz with averaging. BIT 61, 337–359 (2021)
Necoara, I.: Faster randomized block Kaczmarz algorithm. SIAM J. Matrix Anal. Appl. 40, 1425–1452 (2019)
Du, K., Si, W.-T., Sun, X.-H.: Randomized extended average block Kaczmarz for solving least squares. SIAM J. Sci. Comput. 42, A3541–A3559 (2020)
Du, K., Sun, X.-H.: A doubly stochastic block Gauss-Seidel algorithm for solving linear equations. Appl. Math. Comput. 408, 126373 (2021)
Chen, J.-Q., Huang, Z.-D.: On a fast deterministic block Kaczmarz method for solving large-scale linear systems. Numer. Algorithms 89, 1007–1029 (2022)
Motzkin, T.S., Schoenberg, I.J.: The relaxation method for linear inequalities. Can. J. Math. 6, 393–404 (1954)
Bai, Z.-Z., Wu, W.-T.: On greedy randomized Kaczmarz method for solving large sparse linear systems. SIAM J. Sci. Comput. 40, A592–A606 (2018)
Nutini, J.: Greed is good: greedy optimization methods for large-scale structured problems. PhD thesis, University of British Columbia (2018)
De Loera, J.A., Haddock, J., Needell, D.: A sampling Kaczmarz–Motzkin algorithm for linear feasibility. SIAM J. Sci. Comput. 39, S66–S87 (2017)
Morshed, M.S., Islam, M.S., Noor-E-Alam, M.: Accelerated sampling Kaczmarz Motzkin algorithm for the linear feasibility problem. J. Glob. Optim. 77, 361–382 (2020)
Morshed, M.S., Islam, M.S., Noor-E-Alam, M.: Sampling Kaczmarz Motzkin method for linear feasibility problems: generalization and acceleration. Math. Progr. 194, 719–779 (2022)
Liu, Y., Gu, C.-Q.: On greedy randomized block Kaczmarz method for consistent linear systems. Linear Algebra Appl. 616, 178–200 (2021)
Zhang, Y.-J., Li, H.-Y.: Block sampling Kaczmarz–Motzkin methods for consistent linear systems. Calcolo 58, 39 (2021)
Kelley, C.T.: Solving Nonlinear Equations with Newton’s Method. SIAM, Philadelphia (2003)
More, J.J., Garbow, B.S., Hillstrom, K.E.: Testing unconstrained optimization software. ACM Trans. Math. Softw. 7, 17–41 (1981)
Gomes-Ruggiero, M.A., Martänez, J.M., Moretti, A.C.: Comparing algorithms for solving sparse nonlinear systems of equations. SIAM J. Sci. Stat. Comput. 13, 459–483 (1992)
Lukšan, L.: Inexact trust region method for large sparse nonlinear least squares. Kybern. Praha 29, 305–324 (1993)
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The work is supported by the National Natural Science Foundation of China under Grant no. 12061009. The work is supported by Jiangxi Provincial Natural Science Foundation under Grant no. 20202BAB201002.
About this article
Cite this article
Zhang, J., Wang, Y. & Zhao, J. On pseudoinverse-free block maximum residual nonlinear Kaczmarz method for solving large-scale nonlinear system of equations. Japan J. Indust. Appl. Math. 41, 637–657 (2024). https://doi.org/10.1007/s13160-023-00620-8
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13160-023-00620-8
Keywords
- Nonlinear system of equations
- Block nonlinear Kaczmarz method
- \(\mu\)-strongly quasi-convexity
- Pseudoinverse-free
- Star-convexity