On the abs-polynomial expansion of piecewise smooth functions

Tom Streubel has observed that for functions in abs-normal form, generalized Taylor expansions of arbitrary order can be generated by algorithmic piecewise differentiation. Abs-normal form means that the real or vector valued function is defined by an evaluation procedure that involves the absolute value function apart from arithmetic operations and times continuously differentiable univariate intrinsic functions. The additive terms in Streubel's expansion are abs-polynomial, i.e. involve neither divisions nor intrinsics. When and where no absolute values occur, Moore's recurrences can be used to propagate univariate Taylor polynomials through the evaluation procedure with a computational effort of , provided all univariate intrinsics are defined as solutions of linear ODEs. This regularity assumption holds for all standard intrinsics, but for irregular elementaries one has to resort to Faa di Bruno's formula, which has exponential complexity in . As already conjectured, we show that the Moore recurrences can be adapted for regular intrinsics to the abs-normal case. Finally, we observe that where the intrinsics are real analytic the expansions can be extended to infinite series that converge absolutely on spherical domains.

[1]  A. Griewank,et al.  Piecewise Polynomial Taylor Expansions—The Generalization of Faà di Bruno’s Formula , 2020 .

[2]  Andreas Griewank,et al.  An algorithm for nonsmooth optimization by successive piecewise linearization , 2019, Math. Program..

[3]  Andreas Griewank,et al.  Characterizing and Testing Subdifferential Regularity in Piecewise Smooth Optimization , 2019, SIAM J. Optim..

[4]  Andreas Griewank,et al.  Relaxing Kink Qualifications and Proving Convergence Rates in Piecewise Smooth Optimization , 2019, SIAM J. Optim..

[5]  Andreas Griewank,et al.  Finite convergence of an active signature method to local minima of piecewise linear functions , 2018, Optim. Methods Softw..

[6]  Manuel Radons,et al.  A note on surjectivity of piecewise affine mappings , 2017, Optim. Lett..

[7]  Andreas Griewank,et al.  Algorithmic differentiation for piecewise smooth functions: a case study for robust optimization , 2018, Optim. Methods Softw..

[8]  Koichi Kubota,et al.  Enumeration of subdifferentials of piecewise linear functions with abs-normal form , 2018, Optim. Methods Softw..

[9]  Andreas Griewank,et al.  Piecewise linear secant approximation via algorithmic piecewise differentiation , 2017, Optim. Methods Softw..

[10]  Andreas Griewank,et al.  Integrating Lipschitzian dynamical systems using piecewise algorithmic differentiation , 2017, Optim. Methods Softw..

[11]  Tom Streubel,et al.  Generic Construction and Efficient Evaluation of Flow Network DAEs and Their Derivatives in the Context of Gas Networks , 2018, OR.

[12]  Andreas Griewank,et al.  First- and second-order optimality conditions for piecewise smooth objective functions , 2016, Optim. Methods Softw..

[13]  Tanja Clees,et al.  Making Network Solvers Globally Convergent , 2016, SIMULTECH.

[14]  Andreas Griewank,et al.  On Lipschitz optimization based on gray-box piecewise linearization , 2016, Math. Program..

[15]  Manuel Radons,et al.  Direct solution of piecewise linear systems , 2016, Theor. Comput. Sci..

[16]  Laurent Hascoët,et al.  Piecewise Linear AD via Source Transformation , 2016 .

[17]  A. Griewank,et al.  Solving piecewise linear systems in abs-normal form , 2015, 1701.00753.

[18]  Andreas Griewank,et al.  On stable piecewise linearization and generalized algorithmic differentiation , 2013, Optim. Methods Softw..

[19]  Andreas Griewank,et al.  Representation and Analysis of Piecewise Linear Functions in Abs-Normal Form , 2013, System Modelling and Optimization.

[20]  S. Scholtes Introduction to Piecewise Differentiable Equations , 2012 .

[21]  Johannes Willkomm,et al.  Introduction to Automatic Differentiation , 2009 .

[22]  S. K. Mishra,et al.  Nonconvex Optimization and Its Applications , 2008 .

[23]  H. Gecol,et al.  The Basic Theory , 2007 .

[24]  Charles E. M. Pearce,et al.  The multivariate Faà di Bruno formula and multivariate Taylor expansions with explicit integral remainder term , 2007, The ANZIAM Journal.

[25]  F. Giannessi Variational Analysis and Generalized Differentiation , 2006 .

[26]  Diethard Klatte,et al.  Nonsmooth Equations in Optimization: "Regularity, Calculus, Methods And Applications" , 2006 .

[27]  B. Mordukhovich Variational Analysis and Generalized Differentiation II: Applications , 2006 .

[28]  Andreas Griewank,et al.  Introduction to Automatic Differentiation , 2003 .

[29]  D. Klatte Nonsmooth equations in optimization , 2002 .

[30]  V. F. Demyanov,et al.  An Introduction to Quasidifferential Calculus , 2000 .

[31]  Andreas Griewank,et al.  Evaluating derivatives - principles and techniques of algorithmic differentiation, Second Edition , 2000, Frontiers in applied mathematics.

[32]  A. Griewank Automatic Directional Differentiation of Nonsmooth Composite Functions , 1995 .

[33]  F. Clarke Optimization And Nonsmooth Analysis , 1983 .

[34]  Charles M. Whish XXXIII. On the Hindú Quadrature of the Circle, and the infinite Series of the proportion of the circumference to the diameter exhibited in the four S'ástras, the Tantra Sangraham, Yucti Bháshá, Carana Padhati, and Sadratnamála , 1834 .