Abstract
Many problems in science and engineering must solve nonlinear necessary conditions. For example, a standard problem in optimization involves solving for the roots of nonlinear functions defined by f(x) = 0, where x is the unknown variable. Classically one develops a first-order Taylor series model that defines the necessary condition that must be iteratively refined. The standard assumption is that the correction terms are small. Two classes of problems arise: (1) non-square systems that lead to least-squares solutions, and (2) square systems that are often handled by Newton-like methods. The accuracy of the starting guess impacts the number of iteration cycles required. To handle more nonlinear problems, both approaches are generalized to account for first- through fourth-order approximations. Computational differentiation tools are used for automatically formulating and numerically computing the partial derivatives. Two solution approaches are presented for inverting the tensor-valued necessary condition: (1) an integrated Legendre transformation, homotopy method, and high-order vector reversion of series algorithm; and (2) a computational differentiation-based generalized linear algebra approach. Several numerical examples are presented to demonstrate generalized multilinear Least-Squares and Newton-Raphson Methods. Accelerated convergence rates are demonstrated for scalar and vector root-solving problems. The integration of generalized algorithms and automatic differentiation is expected to have broad potential for impacting the design and use of mathematical programming tools for knowledge discovery applications in science and engineering.
Similar content being viewed by others
References
Wengert, R.E.: A simple automatic derivative evaluation program. Comm. AGM 7, 463–464 (1964)
Wilkins, R.D.: Investigation of a new analytical method for numerical derivative evaluation. Comm. ACM 7, 465–471 (1964)
Griewank, A.: On automatic differentiation. In: Iri, M., Tanabe, K. (eds.) Mathematical programming: recent developments and applications, pp. 83–108. Kluwer Academic Publishers, Amsterdam (1989)
Bischof, C., Carle, A., Corliss, G., Griewank, A., Hovland, P.: ADIFOR: generating derivative codes from fortran programs. Sci. Program. 1, 1–29 (1992)
Bishchof, C., Carle, A., Khademi, P., Mauer, A., Hovland, P.: ADIFOR 2.0 User’s Guide Revision CO, “Technical Report ANL/MCS-TM-192, Mathematics and computer Science Division, Argonne National Laboratory, Argonne, IL. (1995)
Eberhard, P., Bischof, C.: Automatic Differentiation of Numerical Integration Algorithms, Technical Report ANL/MCS-P621-1196, Mathematics and Computer Science Division, Argonne National Laboratory, Argonne, IL (1996)
Turner, J.D.: Automated generation of high-order partial derivative models. AIAA J. 41(8), 1590–1599 (2003)
Turner, J.D.: Generalized gradient search and Newton’s method for multilinear algebra root-solving and optimization applications. In: Vadali, S. R., Mortari, D. (eds.) Paper No. AAS-03-261, Proceedings of the John L. Junkins Astrodynamics Symposium 2003, Vol. 115, Advances in the Astronautical Sciences, pp. 55–78. George Bush Conference Center, College Station (2003)
Griffith, D.T., Turner, J.D., Junkins, J.L.: Automatic generation and integration of equation of motion for flexible multibody dynamical systems. AAS J. Astronaut. Sci. 53(3), 251–279 (2005)
Bani Younes, A. H., Turner, J.D., Majji, M., Junkins, J.L.: An Investigation of State Feedback Gain Sensitivity Calculations, Presented to AIAA/AAS Astrodynamics Specialist Conf. Held 2-5 August 2010, Toronto, Ontario, Canada
Griffith, D.T., Turner, J.D., Junkins, J.L.: An embedded function tool for modeling and simulating estimation problems in aerospace engineering, AAS Journal of the Astronautical Sciences, AAS 04-148 (2004)
Griffith, D.J., Turner, J.D., Junkins, J.L.: Some Applications of Automatic Differentiation to Rigid, Flexible, and Constrained Multibody Dynamics, Paper No. DETC2005-85640, Proceedings of the 5th International Conference on Multibody Systems, Nonlinear Dynamics and Control At The ASME 2005 International Design Engineering Technical Conferences, Symposia Title: MSNDC-7-1 Symposium on Applications of Automatic Differentiation for Nonlinear Multibody Dynamics, Control, and Optimization, Long Beach, California, September 24–28 (2005)
Junkins, J.L., Turner, J.D., Majji, M.: Generalizations and applications of the lagrange implicit function theorem, Special issue: The F. Landis Markley Astronautics Symposium. J. Astronaut. Sci. 57(1 and 2), 313–345 (2009)
Majji, M., Junkins, J.L., Turner, J.D.: A Perturbation Method For Estimation Of Dynamic Systems, Nonlinear Dynamics, Online First (2009)
Gauss, C.F.: (1809) Theoria Motus Corporum Coelestium. Perthes, Hamburg. Translation reprinted as Theory of the Motions of the Heavenly bodies moving about the Sun in JConic Sections. Dover, New York (1963)
Crassidis, J.L., Junkins, J.L.: Optimal estimation of dynamical systems. Chapman & Hall/CRC, Boca Raton (2004)
Griffith, T. D.: New methods for estimation, modeling and validation of dynamical systems using automatic differentiation, Ph.D. Dissertation. Texas A&M University, Department of Aerospace Engineering (2004)
Wang, L., Leblanc, A.: Second-order nonlinear least squares estimation. Ann. Inst Stat. Math. 60, 883–900
Wu, C.-F.: Asymptotic theory of nonlinear least squares estimation. Ann. Stat. 9(3), 501–513 (1981)
Zhao, Q., Caiafa, C. F., Mandic, D. P., Chao, Z. C., Nagasaka, Y., Fujii, N., Zhang, L., Cichocki, A.: Higher-Order Partial Least Squares (HOPLS): A Generalized Multi-Linear Regression Method. Cornell University Library. arXiv: 1207.1230 [cs.AL]
Bellman, R. E.: Perturbation Techniques in Mathematics. Engineering & Physics, Holt, Rinehart and Winston, Inc., New York (1964). Part 1, sections 2-5
Liao, S. J.: An approximate solution technique not depending on small parameters: A special example. Int. J. Nonlinear Mech. 30(3), 371–380 (1995)
Newton, Methodus fluxionum et serierum infinitarum, 1664–1671
Raphson, J.: Analysis Aequationum universalis, London (1690)
Kantorovich, L. V., Akilov, G. P.: Functional analysis in normed spaces. Pergamon Press, New York (1982)
Ortega, J. M., Rheinboldt, W. C.: Iterative solution of nonlinear equations in several variables (1970)
Householder, A. S.: The numerical treatment of a single nonlinear equation. McGraw-Hill, New York (1970)
Tranter, C.J.: Legendre transforms. Q. J. Math. 1(1), 1–8. doi:10.1093/qmath/1.1.1,1950
Author information
Authors and Affiliations
Corresponding author
Appendix: Legendre Transformation for Developing a Reversion of Series Solution
Appendix: Legendre Transformation for Developing a Reversion of Series Solution
All of the nonlinear necessary conditions developed in this work lead to n-tuple sets of algebraic equations. The challenge is that these variables fundamentally lack an independent variable, which can be used for developing an analytic continuation solution for the approximate root solution. An artificial independent variable is introduced into the n-tuple algebraic equations by invoking the following Legendre transformation [28]:
where s∈[1, 0] is the artificial independent variable and H represents a homotopy embedding equation. The equation implies that x = x(s). Repeated differentiation of Eq. A.1, with respect to s, yields the cascade of sensitivity necessary conditions
where ∇x,∇2 x,⋯,∇m x are generated by OCEA, and the unknown rates are presented in the red font. Inverting cascade for the rates yields
The solution for x is now analytically continued to fourth order, as a function of the artificial independent variable s, as
Recalling the homotopy parameters defined in Eq. A.1, it follows that the change in the s-variable is
which permits (A.4) to be expressed as the reversion-of-series solution given by
Equation A.6, when combined with Eqs. A.3 and A.4, makes full use of the sensitivity data recovered by the generalized Matrix-vector solution of Eq. 16. A significant advantage of this approach is that the reversion-of-series approach does not depend on the algebraic complexity of the original nonlinear problem, as well as scaling to arbitrary order expansions.
Rights and permissions
About this article
Cite this article
Younes, A.B., Turner, J. Generalized Least Squares and Newton’s Method Algorithms for Nonlinear Root-Solving Applications. J of Astronaut Sci 60, 517–540 (2013). https://doi.org/10.1007/s40295-015-0071-z
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40295-015-0071-z