A new algorithm using L-BFGS for two-parameter eigenvalue problems from Lamé's system and simulation

Year 2024, Volume: 4 Issue: 4, 448 - 468, 30.12.2024

Hayati Ünsal Özer Ahmet Duran

https://doi.org/10.53391/mmnsa.1523702

Abstract

We deal with the challenges and solutions for two-parameter eigenvalue problems (TPEPs) involving large-scale various dense coefficient matrices using several numerical methods. We propose a new method, via fused parameter optimization ($fusedparopt$) algorithm using limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) having several variations, for TPEPs. We combine the advantages of the tensor Rayleigh quotient (RQ), Newton (N) and L-BFGS methods, while avoiding the disadvantages of each method. They are designed for certain TPEPs having real eigenvalue tuples. We test the performance of our algorithm and compare them with state-of-art algorithms such as twopareigs from MultiParEig toolbox in Matlab, tensor Rayleigh quotient-Newton (RQ\_N) and L-BFGS alone, by using the coefficient matrices coming from Lamé system and simulations via randomly generated matrices. We also obtain convergence diagrams for the $fusedparopt\_LBFGS$ to understand the convergence behavior for the number of iterations and computational times via Monte Carlo simulation. We observe that our algorithms can reduce computation time, diminish divergence problems, and give more stable solutions for our data set including various matrices for TPEPs. To the best of our knowledge, we perform the first study including $fusedparopt\_LBFGS$ method in this type of eigenvalue problem.

Keywords

Simulation, Lamé system, numerical linear algebra and matrix theory, algorithm for two-parameter eigenvalue problem, numerical parameter optimization

References

• [1] Atkinson F.V. Multiparameter spectral theory. Bulletin of the American Mathematical Society, 74, 1-27, (1968).
• [2] Atkinson F.V. Multiparameter Eigenvalue Problems: Matrices and Compact Operations. Academic Press: New York, (1972).
• [3] Volkmer, H. Multiparameter Problems and Expansion Theorems. Springer-Verlag: New York, (1988).
• [4] Özer, H.Ü., Duran, A. and Duran, F.S. Fused parameter optimization to find eigenvalue and eigenvectors for three-parameter eigenvalue problems. Submitted.
• [5] Teodorescu, P.P. Treatise on Classical Elasticity. Springer: Dordrecht, (2013).
• [6] Abapolova, E.A. and Soldatov, A.P. Lamé system of elasticity theory in a plane orthotropic medium. Journal of Mathematical Sciences, 157, 387-394, (2009).
• [7] Plestenjak, B., Gheorghiu, C.I. and Hochstenbach, M.E. Spectral collocation for multiparameter eigenvalue problems arising from separable boundary value problems. Journal of Computational Physics, 298, 585-601, (2015).
• [8] Boersma, J. and Jansen, J.K.M. Electromagnetic field singularities at the tip of an elliptic cone. Ph.D. Thesis, Department of Mathematics and Computing Science, Technische Universiteit Eindhoven, (1990).
• [9] Morrison, J.A. and Lewis, J.A. Charge singularity at the corner of a flat plate. SIAM Journal on Applied Mathematics, 31(2), 233-250, (1976).
• [10] Plestenjak, B. MultiParEig MATLAB central file exchange, (Retrieved October 27, 2023). https://www.mathworks.com/matlabcentral/fileexchange/47844-multipareig]
• [11] Bailey, P.B. The automatic solution of two-parameter Sturm-Liouville eigenvalue problems in ordinary differential equations. Applied Mathematics and Computation, 8(4), 251-259, (1981).
• [12] Ji, X. On 2D bisection method for double eigenvalue problems. Journal of Computational Physics, 126(1), 91-98, (1996).
• [13] Jarlebring, E. and Hochstenbach, M.E. Polynomial two-parameter eigenvalue problems and matrix pencil methods for stability of delay-differential equations. Linear Algebra and its Applications, 431(3-4), 369-380, (2009).
• [14] Molzahn, D. Power System Models Formulated as Eigenvalue Problems and Properties of Their Solutions. Ph.D. Thesis, Department of Electrical Engineering, University of Wisconsin-Madison, (2010). [https://minds.wisconsin.edu/handle/1793/46255]
• [15] Plestenjak, B. A continuation method for a right definite two-parameter eigenvalue problem. SIAM Journal on Matrix Analysis and Applications, 21(4), 1163-1184, (2000).
• [16] Hochstenbach, M.E. and Plestenjak B. A Jacobi-Davidson type method for a right definite two-parameter eigenvalue problem. SIAM Journal on Matrix Analysis and Applications, 24(2), 392-410, (2002).
• [17] Hochstenbach, M.E., Kosir, T. and Plestenjak, B. A Jacobi-Davidson type method for the two-parameter eigenvalue problem. SIAM Journal on Matrix Analysis and Applications, 26(2), 477-497, (2004).
• [18] Meerbergen, K. and Plestenjak, B. A Sylvester-Arnoldi type method for the generalized eigenvalue problem with two-by-two operator determinants. Numerical Linear Algebra with Applications, 22(6), 1131-1146, (2015).
• [19] Ringh, E. and Jarlebring, E. Nonlinearizing two-parameter eigenvalue problems. SIAM Journal on Matrix Analysis and Applications, 42(2), 775-799, (2021).
• [20] Blum, E.K. and Geltner, P.B. Numerical solution of eigentuple-eigenvector problems in Hilbert spaces by a gradient method. Numerische Mathematik, 31, 231-246, (1978).
• [21] Blum, E.K. and Curtis, A.R. A convergent gradient method for matrix eigenvector-eigentuple problems. Numerische Mathematik, 31(3), 247-263, (1978).
• [22] Browne, P.J. and Sleeman, B.D. A numerical technique for multiparameter eigenvalue problems. IMA Journal of Numerical Analysis, 2(4), 451–457, (1982).
• [23] Dong, B. The homotopy method for the complete solution of quadratic two-parameter eigenvalue problems. Journal of Scientific Computing, 90, 18, (2022).
• [24] Eisenmann, H. and Nakatsukasa, Y. Solving two-parameter eigenvalue problems using an alternating method. Linear Algebra and its Applications, 643, 137-160, (2022).
• [25] Košir, T. and Plestenjak, B. On the singular two-parameter eigenvalue problem II. Linear Algebra and its Applications, 649, 433-451, (2022).
• [26] Muhiç, A. and Plestenjak, B. On the singular two-parameter eigenvalue problem. Electronic Journal of Linear Algebra, 18, 420-437, (2009).
• [27] Hochstenbach, M.E., Meerbergen, K., Mengi, E. and Plestenjak, B. Subspace methods for three-parameter eigenvalue problems. Numerical Linear Algebra with Applications, 26(4), e2240, (2019).
• [28] Rodriguez, J.I., Du, J.H., You, Y. and Lim, L.H. Fiber product homotopy method for multiparameter eigenvalue problems. Numerische Mathematik, 148, 853-888, (2021).
• [29] Dong, B., Yu, B. and Yu, Y. A homotopy method for finding all solutions of a multiparameter eigenvalue problem. SIAM Journal on Matrix Analysis and Applications, 37(2), 550-571, (2016).
• [30] Trefethen L.N. and Bau III, D. Numerical Linear Algebra. SIAM: Philadelphia, USA, (1997).
• [31] Duran, A. and Caginalp, G. Parameter optimization for differential equations in asset price forecasting. Optimization Methods & Software, 23(4), 551-574, (2008).
• [32] Duran, A. and Tuncel, M. Evaluation of a new parallel numerical parameter optimization algorithm for a dynamical system. In Proceedings, AIP Conference Proceedings, pp. 090052, Pizzo Calabro, Italy, (2016, October).
• [33] Tunçel, M. and Duran, A. Effectiveness of grid and random approaches for a model parameter vector optimization. Journal of Computational Science, 67, 101960, (2023).
• [34] Jackel, P. Monte Carlo Methods in Finance. John Wiley & Sons, England, (2002).
• [35] Kangal, F., Meerbergen, K., Mengi, E. and Michiels, W. A subspace method for large-scale eigenvalue optimization. SIAM Journal on Matrix Analysis and Applications, 39(1), 48-82, (2018).
• [36] Byrd, R.H., Hansen, S.L., Nocedal, J. and Singer, Y. A stochastic quasi-Newton method for large-scale optimization. SIAM Journal on Optimization, 26(2), 1008–1031, (2016).
• [37] Nocedal, J. Updating quasi-Newton matrices with limited storage. Mathematics of Computation, 35(151), 773-782, (1980).
• [38] Saputro, D.R.S. and Widyaningsih, P. Limited memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) method for the parameter estimation on geographically weighted ordinal logistic regression model (GWOLR). In Proceedings, AIP Conference Proceedings, pp. 040009, Yogyakarta, Indonesia, (2017, August).
• [39] Dalvand, Z. and Hajarian, M. Solving generalized inverse eigenvalue problems via L-BFGS-B method. Inverse Problems in Science and Engineering, 28(12), 1719-1746, (2020).
• [40] Nocedal, J. and Wright, S.J. Numerical Optimization. Second Edition, Springer: New York, (2006).
• [41] Peressini, A.L., Sullivan, F.E. and Uhl Jr., J.J. The Mathematics of Nonlinear Programming. Springer New York: New York, (1988).
• [42] Freund, R.M. The Steepest Descent Algorithm for Unconstrained Optimization and a Bisection Line-search Method. Ph.D. Thesis, Department of Electrical Engineering and Computer Science, The University of Massachusetts, (2004). [https://ocw.mit.edu/courses/15-084j-nonlinearprogramming-spring-2004/resources/lec5_steep_desce/]
• [43] Burden, R.L. and Faires, J.D. Numerical Analysis. Brooks/Cole Cengage Learning: Boston, (2011).
• [44] Boyd, S. and Vandenberghe, L. Convex Optimization. Cambridge University Press: Cambridge, (2004).
• [45] Barzilai, J. and Borwein, J.M. Two-point step size gradient methods. IMA Journal of Numerical Analysis, 8(1), 141-148, (1988).
• [46] Horn, R.A. and Johnson, C.R. Topics in Matrix Analysis. Cambridge University Press: Cambridge, (1991).
• [47] Gil, M. Bounds for the spectrum of a two parameter matrix eigenvalue problem. Linear Algebra and its Applications, 498, 201-218 (2016).
• [48] Watkins, D.S. Fundamentals of Matrix Computations. Wiley-Interscience, A John Wiley & Sons: New York, (2002).
• [49] Jonsson, M. An Introduction to Monte Carlo Simulations. Textbook for Numerical Methods with Financial Applications, Department of Mathematics, University of Michigan, (2006).
• [50] June, L.W. and Abu Hassan, M. Modifications of the limited memory bfgs algorithm for large-scale nonlinear optimization. Mathematical Journal of Okayama University, 47(1), 175-188, (2005).
• [51] Mokhtari, A. and Ribeiro, A. Global convergence of online limited memory BFGS. Journal of Machine Learning Research, 16(98), 3151-3181, (2015).