Algorithmic Lie Symmetry Analysis and Group Classication for Ordinary Dierential Equations

Similar documents
Classification of cubics up to affine transformations

On the Linearization of Second-Order Dif ferential and Dif ference Equations

Conservation laws for the geodesic equations of the canonical connection on Lie groups in dimensions two and three

Lie Theory of Differential Equations and Computer Algebra

z = f (x; y) f (x ; y ) f (x; y) f (x; y )

Chapter 3 Second Order Linear Equations

Symmetry Methods for Differential and Difference Equations. Peter Hydon

Chapter 11. Taylor Series. Josef Leydold Mathematical Methods WS 2018/19 11 Taylor Series 1 / 27

Computing the Monodromy Group of a Plane Algebraic Curve Using a New Numerical-modular Newton-Puiseux Algorithm

On the classification of certain curves up to projective tranformations

FINAL EXAM MATH303 Theory of Ordinary Differential Equations. Spring dx dt = x + 3y dy dt = x y.

Notes for Lecture 10

Entrance Exam, Real Analysis September 1, 2017 Solve exactly 6 out of the 8 problems

Symmetry Reduction of Chazy Equation

Alexei F. Cheviakov. University of Saskatchewan, Saskatoon, Canada. INPL seminar June 09, 2011

Chapter 4: First-order differential equations. Similarity and Transport Phenomena in Fluid Dynamics Christophe Ancey

Our goal is to solve a general constant coecient linear second order. this way but that will not always happen). Once we have y 1, it will always

LECTURE 16: LIE GROUPS AND THEIR LIE ALGEBRAS. 1. Lie groups

Department of Mathematics Luleå University of Technology, S Luleå, Sweden. Abstract

Minimum and maximum values *

Symmetry Methods for Differential Equations and Conservation Laws. Peter J. Olver University of Minnesota

MA22S3 Summary Sheet: Ordinary Differential Equations

Ordinary Differential Equations (ODEs)

Math Review ECON 300: Spring 2014 Benjamin A. Jones MATH/CALCULUS REVIEW

Bessel s Equation. MATH 365 Ordinary Differential Equations. J. Robert Buchanan. Fall Department of Mathematics

Degree Master of Science in Mathematical Modelling and Scientific Computing Mathematical Methods I Thursday, 12th January 2012, 9:30 a.m.- 11:30 a.m.

Series Solution of Linear Ordinary Differential Equations

Math 1270 Honors ODE I Fall, 2008 Class notes # 14. x 0 = F (x; y) y 0 = G (x; y) u 0 = au + bv = cu + dv

Type II hidden symmetries of the Laplace equation

1 Ordinary points and singular points

B.7 Lie Groups and Differential Equations

Application of Lie-group symmetry analysis to an infinite hierarchy of differential equations at the example of first order ODEs

7.4* General logarithmic and exponential functions

Chapter 8. Numerical Solution of Ordinary Differential Equations. Module No. 1. Runge-Kutta Methods

Lecture 2: Review of Prerequisites. Table of contents

Linearization of Second-Order Ordinary Dif ferential Equations by Generalized Sundman Transformations

Conditional symmetries of the equations of mathematical physics

Convergence of sequences and series

Conservation Laws: Systematic Construction, Noether s Theorem, Applications, and Symbolic Computations.

arxiv: v1 [math.ap] 25 Aug 2009

(0, 0), (1, ), (2, ), (3, ), (4, ), (5, ), (6, ).

TEST CODE: MMA (Objective type) 2015 SYLLABUS

Lecture Notes on Partial Dierential Equations (PDE)/ MaSc 221+MaSc 225

Introduction and mathematical preliminaries

Math Exam 2, October 14, 2008

Research Article Equivalent Lagrangians: Generalization, Transformation Maps, and Applications

(3) Let Y be a totally bounded subset of a metric space X. Then the closure Y of Y

Solving Differential Equations by Lie Groups. Erol Kavvas

MODELLING OF FLEXIBLE MECHANICAL SYSTEMS THROUGH APPROXIMATED EIGENFUNCTIONS L. Menini A. Tornambe L. Zaccarian Dip. Informatica, Sistemi e Produzione

Math 4263 Homework Set 1

1 Arithmetic calculations (calculator is not allowed)

SYSTEMS OF ODES. mg sin ( (x)) dx 2 =

x 9 or x > 10 Name: Class: Date: 1 How many natural numbers are between 1.5 and 4.5 on the number line?

4. Higher Order Linear DEs

3.1 Basic properties of real numbers - continuation Inmum and supremum of a set of real numbers

Solutions Preliminary Examination in Numerical Analysis January, 2017

14.7: Maxima and Minima

SYMMETRY REDUCTION AND NUMERICAL SOLUTION OF A THIRD-ORDER ODE FROM THIN FILM FLOW

Inductive and Recursive Moving Frames for Lie Pseudo-Groups

Vector Space Basics. 1 Abstract Vector Spaces. 1. (commutativity of vector addition) u + v = v + u. 2. (associativity of vector addition)

Lecture 42 Determining Internal Node Values

Lie Groups for 2D and 3D Transformations

EXAMPLE OF ONE-STEP METHOD

Sophus Lie s Approach to Differential Equations

REVIEW OF DIFFERENTIAL CALCULUS

APPLICATIONS OF FD APPROXIMATIONS FOR SOLVING ORDINARY DIFFERENTIAL EQUATIONS

Math 10C - Fall Final Exam

Sharp Gårding inequality on compact Lie groups.

Math 322. Spring 2015 Review Problems for Midterm 2

Linearization of Two Dimensional Complex-Linearizable Systems of Second Order Ordinary Differential Equations

The following can also be obtained from this WWW address: the papers [8, 9], more examples, comments on the implementation and a short description of

Characterization of the Newtonian free particle system in m 2 dependent variables

In this section, mathematical description of the motion of fluid elements moving in a flow field is

From the Heisenberg group to Carnot groups

2tdt 1 y = t2 + C y = which implies C = 1 and the solution is y = 1

On the singularities of non-linear ODEs

Applications of Symmetries and Conservation Laws to the Study of Nonlinear Elasticity Equations

Introduction to Group Theory

Math 131 Exam 2 Spring 2016

FINAL EXAM, MATH 353 SUMMER I 2015

Symmetries, Fields and Particles 2013 Solutions

TEST CODE: MIII (Objective type) 2010 SYLLABUS

The Erwin Schrodinger International Boltzmanngasse 9. Institute for Mathematical Physics A-1090 Wien, Austria

OR MSc Maths Revision Course

Introduction into thermodynamics

Starting from Heat Equation

Logistic Map, Euler & Runge-Kutta Method and Lotka-Volterra Equations

TEST CODE: PMB SYLLABUS

Ordinary Differential Equations I

LECTURE 15 + C+F. = A 11 x 1x1 +2A 12 x 1x2 + A 22 x 2x2 + B 1 x 1 + B 2 x 2. xi y 2 = ~y 2 (x 1 ;x 2 ) x 2 = ~x 2 (y 1 ;y 2 1

Chapter 3: Transcendental Functions

(1.) For any subset P S we denote by L(P ) the abelian group of integral relations between elements of P, i.e. L(P ) := ker Z P! span Z P S S : For ea

Fourth Order RK-Method

1 + x 2 d dx (sec 1 x) =

Journal of Mathematical Analysis and Applications. Complete symmetry group and nonlocal symmetries for some two-dimensional evolution equations

Initial Value Problems

Logarithmic Functions

Even and odd functions

Ordinary Differential Equations I

Transcription:

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 1 / 25 Algorithmic Lie Symmetry Analysis and Group Classication for Ordinary Dierential Equations Dmitry A. Lyakhov 1 1 Computational Sciences Group, Visual Computing Center King Abdullah University of Science and Technology, Jeddah, Saudi Arabia Kolchin Seminar in Dierential Algebra

Introduction Symmetry group analysis is classical and the most fundamental method for solving dierential equations and constructing solutions with particular properties. Modern Lie theory has a lot of applications in natural sciences: automatic (analytical) solvers of ordinary dierential equations in computer algebra systems construction of conservation laws, equivalence mappings for partial dierential systems numerical (Lie) integrators (which preserve symmetries) reveal much better numerical behavior pattern recognition in computer vision (based on group equivalence of boundary curves) dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 2 / 25

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 3 / 25 Two equations 1. Constant coecient equation 2. Euler equation d n y dx n + a d n 1 y n 1 dx n 1 +... + a dy 1 dx + a 0y = 0. d n y dx n + a n 1x n 1 d n 1 y dx n 1 +... + a 1x dy dx + a 0y = 0.

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 4 / 25 Two equations 1. Constant coecient equation d n y dx n + a d n 1 y n 1 dx n 1 +... + a dy 1 dx + a 0y = 0 admits shifts for x x = x + a and dilations for y ȳ = C y

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 5 / 25 Two equations 2. Euler equation admits dilations for x and dilations for y d n y dx n + a n 1x n 1 d n 1 y dx n 1 +... + a 1x dy dx + a 0y = 0 x = A x ȳ = C y

Two equations Transformation of independent variable, which maps dilation into shift x = A x ln( x) = ln(a) + ln(x) denes change of variables t = exp(x), u = y which maps Euler equation into Constant coecient equation. Remark. Two equivalent equations have isomorphic symmetry groups. dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 6 / 25

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 7 / 25 Symmetry Consider dierential system of general form (x, y (n) ) = 0 (1) where x = (x 1,..., x p ) X and y = (y 1,..., y q ) Y are vectors of independent and dependent variables correspondingly. Denition. The point transformation g : (x, y) ( x, ȳ) of phase space E = X Y is called symmetry if it transforms any solution y(x) of (1) into new function dened by ȳ( x), which is also solution.

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 8 / 25 Symmetry (example) Trivial second-order ODE y (x) = 0

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 9 / 25 Symmetry (example) General solution: y = C 1 x + C 2 Symmetry condition means to transform solutions into solutions, which implies to map straight lines straight lines on plane.

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 10 / 25 Symmetry (example, continue) Trivial second-order ODE y (x) = 0 by substitution [u = f (x, y), t = g(x, y)] is transformed into u (t) + A 3 (u ) 3 + A 2 (u ) 2 + A 1 u + A 0 = 0. Symmetry condition implies A 3 = 0, A 2 = 0, A 1 = 0, A 0 = 0.

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 11 / 25 Symmetry (example, continue) A 3 = 2 g y 2 f y + 2 f y 2 g y = 0, A 2 = 2 g f y 2 x + 2 f g x y y + 2 2 f g y y x 2 2 g f x y y = 0, A 1 = 2 g f x 2 y + 2 f g x 2 y + 2 2 f g x y x 2 2 g f x y x = 0, A 0 = 2 g x 2 f x + 2 f x 2 g x = 0.

Lie symmetry One of the prominent idea of Sophus Lie is to study symmetry properties under one-parameter group of transformation. Denition. Set of transformation g a : (x, y) ( x, ȳ) is called one-parameter group of transformation of dierential system (x, y (n) ) = 0 if 1) g a is symmetry transformation for a, 2) it is a local group: g a g b = g a+b, where (a group parameter). Typically it leads to overdetermined system of linear partial dierential equations of nite type, which could be eciently analyzed symbolically by means of computer algebra for further reduction and explicit solving. dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 12 / 25

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 13 / 25 Lie symmetry (examples) 1. Shift 2. Dilation x = x + a, ȳ = y x = e a x, ȳ = y 3. Rotation ] [ x = ȳ 4. Shear 5. Projection x = [ ] [ ] cos(a) sin(a) x sin(a) cos(a) y x = x + ay, ȳ = y x 1 ax, ȳ = y 1 ax

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 14 / 25 Innitesimal generators The important role plays the rst term in Taylor expansion of one-parameter group of transformation x = x + ε ξ(x, y) + O(ε 2 ), ȳ = y + ε η(x, y) + O(ε 2 ). It is equivalent to And ξ(x, y) = x(x, y, a) a, η(x, y) = a=0 ȳ(x, y, a) a a=0 x (a + b) = Thus assuming b = 0: x(a + b) b = x( x(x, y, a), ȳ(x, y, a), b), ȳ (a + b) =... a x (a) = ξ( x, ȳ), ȳ (a) = η( x, ȳ).

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 15 / 25 Examples 1. Shift 2. Dilation 3. Rotation 4. Shear 5. Projection [ x = ỹ] x = x = x + a, ỹ = y ξ = 1, η = 0 x = e a x, ỹ = y ξ = x, η = 0 [ [ ] cos(a) sin(a) x ξ = y, η = x sin(a) cos(a)] y x = x + ay, ỹ = y ξ = y, η = 0 x 1 ax, ỹ = y 1 ax ξ = x 2, η = xy

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 16 / 25 Symmetry Analysis (example, continue) Substitution to symmetry condition implies General solution g(x, y) := x + a ξ(x, y) + o(a), f (x, y) := y + a η(x, y) + o(a). A 3 = 0, A 2 = 0, A 1 = 0, A 0 = 0 2 η x 2 = 0, 2 ξ x 2 + 2 2 η x y = 0, 2 ξ y 2 = 0, 2 η y 2 + 2 2 ξ x y = 0. ξ(x, y) = (C 7 x + C 8 ) y + C 5 x 2 + C 3 x + C 4, η(x, y) = (C 5 y + C 6 ) x + C 7 y 2 + C 1 y + C 2.

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 17 / 25 Consider ODE (n 2) solved with respect to the highest order derivative where f is rational function. y (n) + f (x, y, y,..., y (n 1) ) = 0, y (k) := d k y dx k (2)

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 18 / 25 Algorithms What can we do algorithmically? generation of determining equations reduction by integrability conditions dimension of solution space structure constants of Lie algebra Initial value problem are given by nite number of values: { l 1 +m 1ξ x l 1 y m 1 (x 0, y 0 ), l 2+m 2η x l 2 y m 2 (x 0, y 0 ) }, [ ξ η ] = n i,j=0 [ ai,j b i,j ] x i y j Since Lie algebra is closed under Lie bracket, then m X = truncated Taylor series [X i, X j ] = Ci,jX k k, 1 i < j m. Remark. Beautiful point of construction that you actually need only truncated Taylor series (approximate solution) to obtain exact values of structure constant. k=1

Linearizability Theorem. Eq. (2) with n 2 is linearizable by a point transformation if and only if one of the following conditions is fullled: 1 n = 2, m = 8; 2 n 3, m = n + 4; 3 n 3, m {n + 1, n + 2} and derived algebra is abelian and has dimension n. Proof. According to group classication of linear equations: 1. Trivial equations y (n) (x) = 0 have maximal possible dimension n + 4 2. Constant coecients equations have Lie algebra span by operators { f 1 (x) y, f 2(x) y,..., f n(x) y, x, y } y 3. Generic case corresponds to Lie algebra { f 1 (x) y, f 2(x) y,..., f n(x) y, y } y dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 19 / 25

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 20 / 25 Group classication The problem of group classication of dierential equations was rst posed by Sophus Lie He also began to solve the problem of group classication of the second-order ordinary equation of general form y + f (x, y, y ) = 0

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 21 / 25 Group classication Lev Ovsyannikov considered simpler case of y + f (x, y) = 0 and solved the problem of group classication by admissible operators f X 1 X 2 X 3 f (y) x 0 0 e y x x x 2 y 0 y k, k 3 x (k 1)x x 2y y 0 ±y 3 x 2x x + y y x 2 x + xy y x 2 g(y) x x 0 0

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 22 / 25 Group classication 2 η x 2 f x ξ f y η + f ( η y 2 ξ x ) = 0, 2 ξ x 2 + 2 2 η ξ 3f x y y = 0, 2 ξ y 2 = 0, 2 η y 2 + 2 2 ξ x y = 0.

dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 23 / 25 Group classication (invariant form) The form of second-order ODE y + f (x, y, y ) = 0 given by f = F 3 (x, y)(y ) 3 + F 2 (x, y)(y ) 2 + F 1 (x, y) y + F 0 (x, y). (3) is invariant under point transformation.

Group classication (linearizable branch) Sophus Lie showed that only equations of the following form are linearizable by point transformations if and only if Theorem. 3(F 3 ) xx 2(F 2 ) xy + (F 1 ) yy 3F 1 (F 3 ) x + 2F 2 (F 2 ) x 3F 3 (F 1 ) x + 3F 0 (F 3 ) y + 6F 3 (F 0 ) y F 2 (F 1 ) y = 0, (4) (F 2 ) xx 2(F 1 ) xy + 3(F 0 ) yy 6F 0 (F 3 ) x + F 1 (F 2 ) x 3F 3 (F 0 ) x + 3F 0 (F 2 ) y + 3F 2 (F 0 ) y 2F 1 (F 1 ) y = 0. dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 24 / 25

Thank you! dmitry.lyakhov@kaust.edu.sa Symbolic Computations May 4, 2018 25 / 25