Modified Equations for Variational Integrators
|
|
- Denis Ferguson
- 5 years ago
- Views:
Transcription
1 Modified Equations for Variational Integrators Mats Vermeeren Technische Universität Berlin Groningen December 18, 2018 Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
2 Table Of Contents 1 Basics of Geometric Numerical Integration [1] 2 Modified Lagrangians [2] 3 Variational integration of the Kepler problem [3] 4 A class of degenerate Lagrangians [4] 5 Variational integrators for Contact Hamiltonian systems [5] 6 Related ideas in integrable systems [6] [1] Hairer, Lubich, Wanner. Geometric Numerical Integration. Springer, [2] V. Modified equations for variational integrators. Numer. Math. 137:1001, [3] V. Numerical precession in variational discretizations of the Kepler problem. In: Discrete Mechanics, Geometric Integration and Lie Butcher Series, Springer, [4] V. Modified equations for variational integrators applied to Lagrangians linear in velocities. arxiv: [5] Work in progress with Alessandro Bravetti and Marcello Seri. [6] V. Continuum limits of Pluri-Lagrangian systems. arxiv: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
3 Table Of Contents 1 Basics of Geometric Numerical Integration [1] 2 Modified Lagrangians [2] 3 Variational integration of the Kepler problem [3] 4 A class of degenerate Lagrangians [4] 5 Variational integrators for Contact Hamiltonian systems [5] 6 Related ideas in integrable systems [6] [1] Hairer, Lubich, Wanner. Geometric Numerical Integration. Springer, [2] V. Modified equations for variational integrators. Numer. Math. 137:1001, [3] V. Numerical precession in variational discretizations of the Kepler problem. In: Discrete Mechanics, Geometric Integration and Lie Butcher Series, Springer, [4] V. Modified equations for variational integrators applied to Lagrangians linear in velocities. arxiv: [5] Work in progress with Alessandro Bravetti and Marcello Seri. [6] V. Continuum limits of Pluri-Lagrangian systems. arxiv: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
4 Continuous Lagrangian Mechanics Lagrange function: L : R 2N = TQ R : (q, q) L(q, q). Solutions are curves q(t) that minimize (or are critical points of) the action S = t1 t 0 L(q(t), q(t)) dt where the integration interval [t 0, t 1 ] and the boundary values q(t 0 ) and q(t 1 ) are fixed. 0 = δs = = t1 t 0 t1 t 0 L L δq + δ q dt q q ( L q d L dt q ) ( ) L t 1 δq dt + q δq t 0 Euler-Lagrange Equation: L q d L dt q = 0. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
5 Continuous Lagrangian Mechanics Lagrange function: L : R 2N = TQ R : (q, q) L(q, q). Solutions are curves q(t) that minimize (or are critical points of) the action S = t1 t 0 L(q(t), q(t)) dt where the integration interval [t 0, t 1 ] and the boundary values q(t 0 ) and q(t 1 ) are fixed. 0 = δs = = t1 t 0 t1 t 0 L L δq + δ q dt q q ( L q d L dt q ) ( ) L t 1 δq dt + q δq t 0 Euler-Lagrange Equation: L q d L dt q = 0. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
6 Legendre transformation Relates Hamiltonian and Lagrangian formalism: Differentiating w.r.t. q, p and q, p = L q q = H p 0 = H q + L q p q = H(q, p) + L(q, q). = ( ) ( H L q + ṗ + q d ) L, dt q establishes equivalence between Lagrangian and Hamiltonian equations of motion. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
7 Legendre transformation Relates Hamiltonian and Lagrangian formalism: Differentiating w.r.t. q, p and q, p = L q q = H p 0 = H q + L q p q = H(q, p) + L(q, q). = ( ) ( H L q + ṗ + q d ) L, dt q establishes equivalence between Lagrangian and Hamiltonian equations of motion. Only works if the Lagrangian is nondegenerate: 2 L q 2 0 Hamiltonian systems preserve the symplectic 2-form ω = i dp i dq i. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
8 Symplectic structure Let Φ t be the flow of a Hamiltonian system, i.e. Φ 0 (q, p) = (q, p) and ( d H dt Φ t(q, p) = p (Φ t(q, p)), H ) q (Φ t(q, p)). Then for each t, Φ t is a symplectic map, Φ t ω = ω, where ω is the canonical symplectic form ω = i dq i dp i. Definition A symplectic integrator is a discretization (in time) of a Hamiltonian systems, such that each discrete time-step is given by a symplectic map. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
9 Discrete Lagrangian mechanics Lagrange function: L : R N R N R : (x, x) L(x, x). Solutions are discrete curves x = (x 0, x 1,..., x n ) that are critical points of the action n S disc = hl disc (x j 1, x j ) Euler-Lagrange equation: j=1 D 2 L disc (x j 1, x j ) + D 1 L disc (x j, x j+1 ) = 0, where D 1, D 2 denote the partial derivatives of L disc. Definition A variational integrator for a continuous system with Lagrangian L is a discrete Lagrangian system with and hence S disc S. L disc ( x(t h), x(t)) L(x(t), ẋ(t)), Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
10 Equivalence Theorem If the Lagrangian/Hamiltonian is regular, variational and symplectic integrators are equivalent. Proof. The discrete Lagrangian is a generation function of the symplectic map describing one time step, p j = hd 1 L disc (x j, x j+1 ) p j+1 = hd 2 L disc (x j, x j+1 ) Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
11 Example: Störmer-Verlet method Consider a mechanical system ẍ = U (x) with Lagrangian L(x, ẋ) = 1 ẋ, ẋ U(x) 2 The Störmer-Verlet discretization is given by the discrete Lagrangian L disc (x j, x j+1 ) = 1 xj+1 x j, x j+1 x j 1 2 h h 2 U (x j) 1 2 U (x j+1) Its discrete Euler-Lagrange equation is x j+1 2x j + x j 1 h 2 = U (x j ) Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
12 Example: Störmer-Verlet method Consider a mechanical system ẍ = U (x) with Lagrangian L(x, ẋ) = 1 ẋ, ẋ U(x) 2 The Störmer-Verlet discretization is given by the discrete Lagrangian L disc (x j, x j+1 ) = 1 xj+1 x j, x j+1 x j 1 2 h h 2 U (x j) 1 2 U (x j+1) Its discrete Euler-Lagrange equation is x j+1 2x j + x j 1 h 2 = U (x j ) Abstract notation: ψ(x j 1, x j, x j+1 ; h) = 0 with ψ(x j 1, x j, x j+1 ; h) = x j+1 2x j + x j 1 h 2 + U (x j ) Symplectic equivalent: x j+1 = x j + hp j h2 2 U (x j ) p j+1 = p j h 2 U (x j ) h 2 U (x j+1 ) Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
13 Modified Equations Exact solution of a differential equation: Numerical solution with a variational integrator: Notice conservation of Energy: Easy to prove for (continuous) Hamiltonian systems Follows by Noether s theorem from invariance under time-translation of the Lagrangian Symplectic/variational integrators very nearly preserve energy. Why? Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
14 Modified Equations Conservation of Energy: Easy to prove for (continuous) Hamiltonian systems Follows by Noether s theorem from invariance under time-translation of the Lagrangian Symplectic/variational integrators very nearly preserve energy. Why? Idea of proof: find a modified equation, a differential equation with solutions that interpolate the numerical solutions: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
15 Modified Equations Modified equations are usually given by power series. Often they do not converge. Definition The differential equation ẍ = f (x, ẋ; h), where f (x, ẋ; h) f 0 (x, ẋ) + hf 1 (x, ẋ) + h 2 f 2 (x, ẋ) +... is a modified equation for the second order difference equation Ψ(x j 1, x j, x j+1 ; h) = 0 if, for every k, every solution of the truncated differential equation ẍ = T k ( fh (x, ẋ) ) = f 0 (x, ẋ) + hf 1 (x, ẋ) + h 2 f 2 (x, ẋ) h k f k (x, ẋ). satisfies Ψ ( x(t h), x(t), x(t + h); h ) = O ( h k+1). Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
16 Modified Equations for symplectic integrators Symplectic integrators are known the very nearly preserve energy, because Theorem The modified equation for a symplectic integrator is a Hamiltonian equation. Can we arrive at a similar result purely on the Lagrangian side? Are modified equations for variational integrators Lagrangian? Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
17 Table Of Contents 1 Basics of Geometric Numerical Integration [1] 2 Modified Lagrangians [2] 3 Variational integration of the Kepler problem [3] 4 A class of degenerate Lagrangians [4] 5 Variational integrators for Contact Hamiltonian systems [5] 6 Related ideas in integrable systems [6] [1] Hairer, Lubich, Wanner. Geometric Numerical Integration. Springer, [2] V. Modified equations for variational integrators. Numer. Math. 137:1001, [3] V. Numerical precession in variational discretizations of the Kepler problem. In: Discrete Mechanics, Geometric Integration and Lie Butcher Series, Springer, [4] V. Modified equations for variational integrators applied to Lagrangians linear in velocities. arxiv: [5] Work in progress with Alessandro Bravetti and Marcello Seri. [6] V. Continuum limits of Pluri-Lagrangian systems. arxiv: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
18 General idea Look for a modified Lagrangian L mod (x, ẋ) such that the discrete Lagrangian L disc is its exact discrete Lagrangian, i.e. jh (j 1)h L mod (x(t), ẋ(t))dt = hl disc ( x((j 1)h), x(jh) ). The Euler-Lagrange equation of L mod will then be the modified equation. The best we can hope for is to find such a modified Lagrangian up to an error of arbitrarily high order in h. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
19 The discrete Lagrangian evaluated on a continuous curve We can write the discrete Lagrangian as a function of x and its derivatives, all evaluated at the point jh h 2, L disc [x] = L disc ( x h 2 ẋ Here and in the following: x + h 2 ẋ = L disc (x j 1, x j ; h) ( ) h 2 ẍ..., 2 ( h 2 ) 2 ẍ +..., h). [x] denotes dependence on x and any number of its derivatives, we evaluate at t = jh h 2 whenever we omit the variable t, i.e. x = x ( jh h 2), x j = x(jh) and x j 1 = x((j 1)h). Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
20 A truly continuous Lagrangian We want to write the discrete action n S disc = hl disc (x j 1, x j ) = as an integral. j=1 Lemma (Euler-MacLaurin formula) n j=1 For any smooth function f : R R N we have n j=1 ( hf jh h ) nh 2 0 = nh 0 where B i are the Bernoulli numbers. [ ( )] hl disc x jh h 2 h 2i ( 2 1 2i 1 ) B 2i (2i)! f (2i) (t) dt i=0 (f (t) h2 24 f (t) + 7h f (4) (t) +... The symbol indicates that this is an asymptotic series. ) dt, Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
21 A truly continuous Lagrangian Definition We call L mesh [x(t)] = L disc [x(t)] + i=1 d 2 ( 2 1 2i 1 ) h 2i B 2i (2i)! d 2i dt 2i L disc[x(t)] = L disc [x(t)] h2 24 dt 2 L disc[x(t)] + 7h dt 4 L disc[x(t)] +... the meshed modified Lagrangian of L disc. Formally, the meshed modified Lagrangian satisfies L mesh [x(t)] dt = hl disc (x j, x j+1 ) where x j = x(jh). d 4 Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
22 A truly continuous Lagrangian Definition We call L mesh [x(t)] = L disc [x(t)] + i=1 d 2 ( 2 1 2i 1 ) h 2i B 2i (2i)! d 2i dt 2i L disc[x(t)] = L disc [x(t)] h2 24 dt 2 L disc[x(t)] + 7h dt 4 L disc[x(t)] +... the meshed modified Lagrangian of L disc. Formally, the meshed modified Lagrangian satisfies L mesh [x(t)] dt = hl disc (x j, x j+1 ) where x j = x(jh). Are we finished? L mesh [x] depends on many more derivatives than the original L(x, ẋ). Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56 d 4
23 The meshed variational problem Definition classical variational problem: find critical curves of some action b a L[x(t)] dt in the set of smooth curves C. meshed variational problem: find critical curves of some action b a L[x(t)] dt in the set of piecewise smooth curves that are consistent with a mesh of size h, C M,h = {x C 0 ([a, b]) t 0 [a, b] : t [a, b] : x not smooth at t t t 0 hn}. x t Classical variational problem x Meshed variational problem t Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
24 The meshed variational problem Criticality conditions of a meshed variational problem: δl Euler-Lagrange equations: δx = 0, where Natural interior conditions: j 2 : k=0 δl = 0, δx (j) or equivalently: j 2 : δl δx (j) = ( 1) k dk L dt k x (j+k). L = 0, x (j) If L is a non-convergent power series, these equations are formal. x t x Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56 t
25 Exploiting the natural interior conditions S disc (x(0), x(h),...) = L mesh ([x], h) dt, hence S disc critical δl mesh δx = 0. Variations that are supported on a single mesh interval do not change the discrete action do not change L mesh ([x], h) dt. are the variations that produce natural interior conditions It follows that for L mesh the NIC are automatically satisfied: δl mesh δx = 0 j 2 : L mesh x (j) = 0 Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
26 Definition The modified Lagrangian is the formal power series L mod (x, ẋ) = L mesh [x] ẍ=fh (x,ẋ), x (3) = d dt f h(x,ẋ),..., where ẍ = f h (x, ẋ) is the modified equation. The k-th truncation of the modified Lagrangian is ) L mod,k = T k (L mod (x, ẋ)) = T k (L mesh [x], x (j) =F j k 2 (x,ẋ) where T k denotes truncation after the h k -term and ẍ = Fk 2 (x, ẋ; h) + O(hk+1 ) = F k (x, ẋ; h) + O(h k+1 ), x (3) = Fk 3 (x, ẋ; h) + O(hk+1 ), x (4) = Fk 4 (x, ẋ; h) + O(hk+1 ),.... are the k-th truncation of the modified equation and its derivatives. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
27 Lemma The meshed modified Lagrangian L mesh [x] and the modified Lagrangian L mod (x, ẋ) have the same critical curves. Proof. L mod,k x = L mesh x + L mesh ẍ Fk 2 x + L mesh F k x (3) x +... x (j) =F j k 1 (x,ẋ) Also, L mod,k ẋ = L mesh x = L mesh ẋ + O(h k+1 ), + L mesh ẍ Fk 2 ẋ + L mesh Fk 3 x (3) ẋ +... x (j) =F j k 1 (x,ẋ) L mod,k x = L mesh ẋ d dt + O(h k+1 ), L mod,k ẋ = j=0 ( 1) j d j dt j L mesh x (j) + O(h k+1 ). Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
28 Main result Theorem For a discrete Lagrangian L disc that is a consistent discretization of some L, the k-th truncation of the Euler-Lagrange equation of L mod,k (x, ẋ) is the k-th truncation of the modified equation. Proof. Let x be a solution of the Euler-Lagrange equation for L mod (x, ẋ). Consider the discrete curve x j = x(jh). x is critical for the action L mod (x, ẋ) dt. By the Lemma, x is critical for the action L mesh [x] dt. By construction, the actions S disc = j L disc(y(jh), y((j + 1)h)) and S = b a L mod[y(t)] dt are equal for any smooth curve y. Therefore the discrete curve (x(jh)) j is critical for the discrete action S disc. Hence D 2 L disc (x(t h), x(t)) + D 1 L disc (x(t), x(t + h)) = 0. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
29 Table Of Contents 1 Basics of Geometric Numerical Integration [1] 2 Modified Lagrangians [2] 3 Variational integration of the Kepler problem [3] 4 A class of degenerate Lagrangians [4] 5 Variational integrators for Contact Hamiltonian systems [5] 6 Related ideas in integrable systems [6] [1] Hairer, Lubich, Wanner. Geometric Numerical Integration. Springer, [2] V. Modified equations for variational integrators. Numer. Math. 137:1001, [3] V. Numerical precession in variational discretizations of the Kepler problem. In: Discrete Mechanics, Geometric Integration and Lie Butcher Series, Springer, [4] V. Modified equations for variational integrators applied to Lagrangians linear in velocities. arxiv: [5] Work in progress with Alessandro Bravetti and Marcello Seri. [6] V. Continuum limits of Pluri-Lagrangian systems. arxiv: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
30 Example: Störmer-Verlet discretization L(x, ẋ) = 1 2 ẋ 2 U(x), L disc (x j, x j+1 ) = 1 x j+1 x j 2 h Its Euler-Lagrange equation is We have L disc [x] = x j+1 2x j + x j 1 h 2 = U (x j ). ẋ + h2 24 x (3) +..., ẋ + h U (x j) 1 2 U (x j+1). 24 x (3) +... ( h 2) 2 ẍ... ) 1 2 U ( x + h 2ẋ (x U h 2ẋ = 1 2 ẋ 2 U + h2 ( ẋ, x (3) 3U ẍ 3U (ẋ, ẋ)) + O(h 4 ) 24 ( h 2) 2 ẍ +... ) Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
31 Example: Störmer-Verlet discretization L disc [x] = 1 2 ẋ 2 U + h2 ( ẋ, x (3) 3U ẍ 3U (ẋ, ẋ)) + O(h 4 ), 24 From this we calculate the meshed modified Lagrangian, d 2 L mesh [x] = L disc [x] h2 24 dt 2 L disc[x] + O(h 4 ) = 1 2 ẋ 2 U + h2 ( ẋ, x (3) 3U ẍ 3U (ẋ, ẋ)) 24 h2 ( ẍ, ẍ + ẋ, x (3) ) U ẍ U (ẋ, ẋ) + O(h 4 ) 24 = 1 2 ẋ 2 U + h2 ( ẍ, ẍ 2U ẍ 2U (ẋ, ẋ) ) + O(h 4 ). 24 Eliminate second derivatives using ẍ = U + O(h 2 ), L mod,3 (x, ẋ) = 1 2 ẋ 2 U + h2 ( U 2 2U (ẋ, ẋ) ). 24 Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
32 Example: Störmer-Verlet discretization The modified Lagrangian is L mod,3 (x, ẋ) = 1 2 h2 ( ẋ, ẋ U + U 2 2U (ẋ, ẋ) ). 24 Observe that this Lagrangian is not separable for general U. The corresponding Euler-Lagrange equation is ẍ U + h2 ( 2U U 2U (ẋ, ẋ) + 4U (ẋ, ẋ) + 4U ẍ ) = Solving this for ẍ we find the modified equation ẍ = U + h2 ( U (ẋ, ẋ) U U ) + O(h 4 ). 12 Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
33 The Kepler problem Potential: U(x) = 1 x. Lagrangian: L = 1 1 ẋ, ẋ + 2 x. Equation of motion ẍ = x x Störmer-Verlet discretization: x j+1 2x j + x j 1 h 2 = U (x j ). 2 3 Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
34 The Kepler problem Potential: U(x) = 1 x. Lagrangian: L = 1 1 ẋ, ẋ + 2 x. Equation of motion ẍ = x x Störmer-Verlet discretization: x j+1 2x j + x j 1 h 2 = U (x j ). Midpoint discretization: x j+1 2x j + x j 1 h 2 = 1 2 U ( xj 1 + x j 2 ) 1 ( ) xj + x j+1 2 U Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
35 Störmer-Verlet discretization of the Kepler problem The modified Lagrangian of the Störmer-Verlet discretization is L mod,3 (x, ẋ) = 1 2 ẋ 2 U + h2 ( U U 2U (ẋ, ẋ) ). 24 For the Kepler problem we have U(x) = 1 x, hence ( ( ) 1 L mod,3 x, ẋ = 2 ẋ ẋ, 2 ) x + h2 1 ẋ x, ẋ 24 x 4 2 x x 5. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
36 Laplace-Runge-Lenz vector The shape and orientation of an orbit for the kepler Problem is determined by the Laplace-Runge-Lenz vector, which is the Noether integral for a generalized variational symmetry. For the modified Lagrangian this is only an approximate symmetry. Lemma (From a perturbative version of Noether s theorem) The precession rate for the perturbed Lagrangian L = 1 2 is in first order approximation 1 ẋ, ẋ + + U(x, ẋ), x 2 U(x, ẋ) 2πa b radians per period, where a and b are the semimajor and semiminor axes, and denotes the time-average along the unperturbed orbit. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
37 Störmer-Verlet discretization of the Kepler problem Proposition The numerical precession rate of the Störmer-Verlet method is π (15 a3 24 b 6 3 a ) b 4 h 2 + O(h 4 ) Predicted: rad per revolution. Measured: rad per revolution. 3 3 Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
38 Midpoint discretization of the Kepler problem Proposition The numerical precession rate of the midpoint rule is π (15 a3 12 b 6 3 a ) b 4 h 2 + O(h 4 ) Predicted: rad per revolution. Measured: rad per revolution. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
39 New methods Precession rate Störmer-Verlet: Precession rate Midpoint rule: π 12 π (15 a3 24 b 6 3 a b 4 (15 a3 b 6 3 a b 4 ) h 2 + O(h 4 ) ) h 2 + O(h 4 ) This allows us to construct new integrators with precession of order h 4. Mixed Lagrangian L(x j, x j+1 ) = 2 3 L SV (x j, x j+1 ) L MP(x j, x j+1 ) Lagrangian composition L j (x k, x k+1 ) = { L MP (x k, x k+1 ) L SV (x k, x k+1 ) if 3 j, otherwise. Composition { of difference equations ( ) ( ) xj+1 2x j + x j 1 = h2 2 U xj 1 +x j 2 h2 2 U xj +x j+1 2 x j+1 2x j + x j 1 = h 2 U (x j ) Is this a variational integrator? if j 2 mod 3, otherwise. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
40 Qualitative analysis of new methods Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
41 Precession rates Precession rate MP SV FR LC ML DEC C h 2 reference h 4 reference MP,SV: old methods LC, ML, DEC: new methods FR: Forest, Ruth. Fourth-order symplectic integration, C: Chin. Symplectic integrators from composite operator factorizations, Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
42 Table Of Contents 1 Basics of Geometric Numerical Integration [1] 2 Modified Lagrangians [2] 3 Variational integration of the Kepler problem [3] 4 A class of degenerate Lagrangians [4] 5 Variational integrators for Contact Hamiltonian systems [5] 6 Related ideas in integrable systems [6] [1] Hairer, Lubich, Wanner. Geometric Numerical Integration. Springer, [2] V. Modified equations for variational integrators. Numer. Math. 137:1001, [3] V. Numerical precession in variational discretizations of the Kepler problem. In: Discrete Mechanics, Geometric Integration and Lie Butcher Series, Springer, [4] V. Modified equations for variational integrators applied to Lagrangians linear in velocities. arxiv: [5] Work in progress with Alessandro Bravetti and Marcello Seri. [6] V. Continuum limits of Pluri-Lagrangian systems. arxiv: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
43 Lagrangians linear in velocities L : T R N = R 2N R of the form L(q, q) = α(q), q H(q), where α : R N R N, H : R N R, and the brackets, denote the standard scalar product. Let ( ) A(q) = α αi (q) (q) = q j i,j=1,...,n and A skew (q) = A(q) T A(q) We assume that A skew (q) is invertible, then the Euler-Lagrange equation for L is given by q = A skew (q) 1 H (q) T Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
44 Examples of Lagrangians linear in velocities Dynamics of point vortices in the (complex) plane L(z, z, ż, ż) = N Γ j Im(z j ż j ) 1 π j=1 ż j = i 2π k j Variational formulation in phase space N j 1 Γ j Γ k log z j z k, j=1 k=1 Γ k z j z k for j = 1,..., N. L(p, q, ṗ, q) = p, q H(p, q). ( ) H T ( ) H T q = and ṗ =. p q Guiding centre motion (plasma physics) Many PDEs, e.g. nonlinear Schrödinger equation. (But modified equations are not so useful for PDEs) Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
45 Possible discretization of L(q, q) = q T A q H(q) ( ) qj + q T ( j+1 qj+1 q j L disc (q j, q j+1, h) = A 2 h ) 1 2 H(q j) 1 2 H(q j+1) Discrete EL equation q j+1 q j 1 2h = (A T A) 1 H (q j ) T. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
46 Possible discretization of L(q, q) = q T A q H(q) ( ) qj + q T ( j+1 qj+1 q j L disc (q j, q j+1, h) = A 2 h Discrete EL equation q j+1 q j 1 2h = (A T A) 1 H (q j ) T. ) 1 2 H(q j) 1 2 H(q j+1) The EL equation involves 3 points needs 2 points of initial data. The differential equation is of 1st order needs only 1 point of initial data. This means we are dealing with a 2-step method. We have a multi-step method, so parasitic oscillations may occur. Solution: double the dimension of the system: principal and parasitic part. The enlarged system is still Lagrangian. We cannot replace q in the Lagrangian. Reason: NIC only involve q, q (3),. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
47 Modified equations for 2-step methods Principal modified equation satisfies q = f (q) + hf 1 (q) + h 2 f 2 (q) h k f k (q) a 0 q(t) + a 1 q(t + h) + a 2 q(t + 2h) h = b 0 f (q(t)) + b 1 f (q(t + h)) + b 2 f (q(t + 2h)) + O(h k+1 ). Full system of modified equations ẋ = f 0 (x, y) + hf 1 (x, y) h k f k (x, y) ẏ = g 0 (x, y) + hg 1 (x, y) h k g k (x, y), such that the discrete curve q j = x(t + jh) + ( 1) j y(t + jh) satisfies a 0 q j + a 1 q j+1 + a 2 q j+2 h = b 0 f (q j ) + b 1 f (q j+1 ) + b 2 f (q j+2 ) + O(h k+1 ) Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
48 Doubling the dimension The discrete curve (x j, y j ) j Z is critical for L(x j, y j, x j+1, y j+1, h) = 1 2 L(x j+y j, x j+1 y j+1, h)+ 1 2 L(x j y j, x j+1 +y j+1, h), if and only if the discrete curves (q + j ) j Z and (q j ) j Z, defined by are critical for L(q j, q j+1, h). q ± j = x j ± ( 1) j y j, Lagrangian for the full system of modified equations = Lagrangian for the principal modified equation of the extended system. Hence we can calculate a Lagrangian for the full system of modified equations with the tools we already have. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
49 Example 1 For L disc (q j, q j+1, h) = we find 1 2 Aq j Aq j+1, q j+1 q j 1 h 2 H(q j) 1 2 H(q j+1) L mod,0 (x, y, ẋ, ẏ, h) = Ax, ẋ + Aẏ, y 1 2 H(x + y) 1 H(x y). 2 Its Euler-Lagrange equations are ( 1 ẋ = A 1 skew 2 H (x + y) T + 1 ) 2 H (x y) T + O(h), ẏ = A 1 skew ( 12 H (x + y) T + 12 ) H (x y) T + O(h). Linearize the second equation around y = 0 ẏ = A 1 skew H (x)y + O( y 2 + h) Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
50 Example 1 Magnitude of oscillations satisfies ẏ = A 1 skew H (x)y + O( y 2 + h) Unless the matrix A 1 skew H (x) is exceptionally friendly, we expect growing parasitic oscillations. (Note that an eigenvalue analysis does not apply because A 1 skew H (x) is not constant) Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
51 Example 2 For we find L disc (q j, q j+1, h) = A q j + q j+1, q ( ) j+1 q j qj + q j+1 H 2 h 2 L mod,0 (x, y, ẋ, ẏ, h) = Ax, ẋ + Aẏ, y H(x). Its Euler-Lagrange equations are ẋ = A 1 skew H (x) T + O(h), ẏ = 0 + O(h). Even better, ẏ = 0 to any order no growing oscillations Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
52 Table Of Contents 1 Basics of Geometric Numerical Integration [1] 2 Modified Lagrangians [2] 3 Variational integration of the Kepler problem [3] 4 A class of degenerate Lagrangians [4] 5 Variational integrators for Contact Hamiltonian systems [5] 6 Related ideas in integrable systems [6] [1] Hairer, Lubich, Wanner. Geometric Numerical Integration. Springer, [2] V. Modified equations for variational integrators. Numer. Math. 137:1001, [3] V. Numerical precession in variational discretizations of the Kepler problem. In: Discrete Mechanics, Geometric Integration and Lie Butcher Series, Springer, [4] V. Modified equations for variational integrators applied to Lagrangians linear in velocities. arxiv: [5] Work in progress with Alessandro Bravetti and Marcello Seri. [6] V. Continuum limits of Pluri-Lagrangian systems. arxiv: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
53 Contact Hamiltonian dynamics Contact geometry is an odd-dimensional counterpart to symplectic geometry. In mechanics, we typically work on T Q R, a Hamilton function H : T Q R R : (q, p, z) H(q, p, z) generates dynamics by q = H p ṗ = H q p H z ż = p H p H Example. system: H(q, p, z) = 1 2 p2 + V (q) + αz describes a damped mechanical q = p ṗ = V (q) αp ż = p 2 H Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
54 Contact variational principle Given a Lagrange function consider the ode L : TQ R R, Given a discrete Lagrange function L : Q Q R R, consider the ode ż = L(q, q, z) on a fixed interval [0, T ], with z(0), q(0) and q(t ) prescribed. Herglotz variational principle: z(t ) critical under variations of the curve q. z j+1 z j h = L(q j, q j+1, z j ) for j {0,..., N} with z 0, q 0 and q N prescribed. Discrete Herglotz variational principle: z N critical under variations of the discrete curve q. In this framework, we can adapt the theory of variational integrators to contact systems. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
55 Table Of Contents 1 Basics of Geometric Numerical Integration [1] 2 Modified Lagrangians [2] 3 Variational integration of the Kepler problem [3] 4 A class of degenerate Lagrangians [4] 5 Variational integrators for Contact Hamiltonian systems [5] 6 Related ideas in integrable systems [6] [1] Hairer, Lubich, Wanner. Geometric Numerical Integration. Springer, [2] V. Modified equations for variational integrators. Numer. Math. 137:1001, [3] V. Numerical precession in variational discretizations of the Kepler problem. In: Discrete Mechanics, Geometric Integration and Lie Butcher Series, Springer, [4] V. Modified equations for variational integrators applied to Lagrangians linear in velocities. arxiv: [5] Work in progress with Alessandro Bravetti and Marcello Seri. [6] V. Continuum limits of Pluri-Lagrangian systems. arxiv: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
56 Integrable systems An integrable system is (a system of) nonlinear differential or difference equation(s), that to some extend behaves like a linear system. Our perspective An equation is integrable if it is part of a sufficiently large system of compatible equations. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
57 Integrable systems An integrable system is (a system of) nonlinear differential or difference equation(s), that to some extend behaves like a linear system. Our perspective An equation is integrable if it is part of a sufficiently large system of compatible equations. In Mechanics: A Hamiltonian system with Hamilton function H : T Q R 2N R is Liouville-Arnold integrable if there exist N functionally independent Hamilton functions H = H 1, H 2,... H N in involution: {H i, H j } = 0. (1+1)-dimensional PDEs: Infinite sequence H = H 1, H 2,... of Hamiltonians in involution. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
58 Integrable systems An integrable system is (a system of) nonlinear differential or difference equation(s), that to some extend behaves like a linear system. Our perspective An equation is integrable if it is part of a sufficiently large system of compatible equations. In Mechanics: A Hamiltonian system with Hamilton function H : T Q R 2N R is Liouville-Arnold integrable if there exist N functionally independent Hamilton functions H = H 1, H 2,... H N in involution: {H i, H j } = 0. (1+1)-dimensional PDEs: Infinite sequence H = H 1, H 2,... of Hamiltonians in involution. Variational description of an integrable hierarchy: pluri-lagrangian or Lagrangian multiform systems. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
59 Quad equations Quad equation: U 2 U 12 Q(U, U 1, U 2, U 12, α 1, α 2 ) = 0 α 2 Subscripts of U denote lattice shifts, α 1, α 2 are parameters. Invariant under symmetries of the square, affine in each of U, U 1, U 2, U 12. U U 23 α 1 U 1 U 123 Integrability for systems quad equations: Multi-dimensional consistency of U 3 U 13 Q(U, U i, U j, U ij, α i, α j ) = 0, α 3 U 2 i.e. the three ways of calculating U 123 give the same result. U α 2 α 1 U 1 U 12 Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
60 Pluri-Lagrangian structure for quad equations For some discrete 2-form L( ij ) = L(U, U i, U j, U ij, α i, α j ), the action L( ) is critical on all 2-surfaces Γ in N N simultaneously. Γ [Lobb, Nijhoff. Lagrangian multiforms and multidimensional consistency. J. Phys. A. 2009] Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
61 Continuous Pluri-Lagrangian systems A field u : R N Q is a solution of the pluri-lagrangian problem for the Lagrangian 2-form, L = L ij [u] dt i dt j. i,j if the action Γ L is critical on all smooth surfaces Γ in RN. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
62 Continuous Pluri-Lagrangian systems A field u : R N Q is a solution of the pluri-lagrangian problem for the Lagrangian 2-form, L = L ij [u] dt i dt j. i,j if the action Γ L is critical on all smooth surfaces Γ in RN. Continuum limits of pluri-lagrangian systems use many of the ideas presented today. Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
63 Summary 1 Basics of Geometric Numerical Integration [1] 2 Modified Lagrangians [2] 3 Variational integration of the Kepler problem [3] 4 A class of degenerate Lagrangians [4] 5 Variational integrators for Contact Hamiltonian systems [5] 6 Related ideas in integrable systems [6] [1] Hairer, Lubich, Wanner. Geometric Numerical Integration. Springer, [2] V. Modified equations for variational integrators. Numer. Math. 137:1001, [3] V. Numerical precession in variational discretizations of the Kepler problem. In: Discrete Mechanics, Geometric Integration and Lie Butcher Series, Springer, [4] V. Modified equations for variational integrators applied to Lagrangians linear in velocities. arxiv: [5] Work in progress with Alessandro Bravetti and Marcello Seri. [6] V. Continuum limits of Pluri-Lagrangian systems. arxiv: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
64 Summary Obtaining a high-order modified Lagrangian L mesh [x] is relatively straightforward, but its interpretation is not. From L mesh [x] a first order Lagrangian L mod,k (x, ẋ) can be found using the meshed variational principle. (This might be simpler in the contact formulation.) If the Lagrangian is nondegenerate, the modified Lagrangian can also be obtained by Legendre transform from the modified Hamiltonian. Our approach extends to degenerate Lagrangians that are linear in velocities. Similar ideas apply to continuum limits of integrable systems. Can we get improved error estimates from the Lagrangian perspective? What about nonholonomic constraints? What about PDEs? Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
65 Thank you for your attention! 1 Basics of Geometric Numerical Integration [1] 2 Modified Lagrangians [2] 3 Variational integration of the Kepler problem [3] 4 A class of degenerate Lagrangians [4] 5 Variational integrators for Contact Hamiltonian systems [5] 6 Related ideas in integrable systems [6] [1] Hairer, Lubich, Wanner. Geometric Numerical Integration. Springer, [2] V. Modified equations for variational integrators. Numer. Math. 137:1001, [3] V. Numerical precession in variational discretizations of the Kepler problem. In: Discrete Mechanics, Geometric Integration and Lie Butcher Series, Springer, [4] V. Modified equations for variational integrators applied to Lagrangians linear in velocities. arxiv: [5] Work in progress with Alessandro Bravetti and Marcello Seri. [6] V. Continuum limits of Pluri-Lagrangian systems. arxiv: Mats Vermeeren (TU Berlin) Modified equations for variational integrators December 18, / 56
Backward error analysis
Backward error analysis Brynjulf Owren July 28, 2015 Introduction. The main source for these notes is the monograph by Hairer, Lubich and Wanner [2]. Consider ODEs in R d of the form ẏ = f(y), y(0) = y
More informationSymmetric multistep methods for charged-particle dynamics
Version of 12 September 2017 Symmetric multistep methods for charged-particle dynamics Ernst Hairer 1, Christian Lubich 2 Abstract A class of explicit symmetric multistep methods is proposed for integrating
More informationSymplectic integration with Runge-Kutta methods, AARMS summer school 2015
Symplectic integration with Runge-Kutta methods, AARMS summer school 2015 Elena Celledoni July 13, 2015 1 Hamiltonian systems and their properties We consider a Hamiltonian system in the form ẏ = J H(y)
More informationHamiltonian Dynamics
Hamiltonian Dynamics CDS 140b Joris Vankerschaver jv@caltech.edu CDS Feb. 10, 2009 Joris Vankerschaver (CDS) Hamiltonian Dynamics Feb. 10, 2009 1 / 31 Outline 1. Introductory concepts; 2. Poisson brackets;
More informationChapter 1 Symplectic Integrator and Beam Dynamics Simulations
Chapter 1 and Beam Accelerator Physics Group, Journal Club November 9, 2010 and Beam NSLS-II Brookhaven National Laboratory 1.1 (70) Big Picture for Numerical Accelerator Physics and Beam For single particle
More informationGeometric Mechanics and Global Nonlinear Control for Multi-Body Dynamics
Geometric Mechanics and Global Nonlinear Control for Multi-Body Dynamics Harris McClamroch Aerospace Engineering, University of Michigan Joint work with Taeyoung Lee (George Washington University) Melvin
More information1 Ordinary differential equations
Numerical Analysis Seminar Frühjahrssemester 08 Lecturers: Prof. M. Torrilhon, Prof. D. Kressner The Störmer-Verlet method F. Crivelli (flcrivel@student.ethz.ch May 8, 2008 Introduction During this talk
More informationChapter 1. Principles of Motion in Invariantive Mechanics
Chapter 1 Principles of Motion in Invariantive Mechanics 1.1. The Euler-Lagrange and Hamilton s equations obtained by means of exterior forms Let L = L(q 1,q 2,...,q n, q 1, q 2,..., q n,t) L(q, q,t) (1.1)
More informationIntroduction to Hamiltonian systems
Introduction to Hamiltonian systems Marlis Hochbruck Heinrich-Heine Universität Düsseldorf Oberwolfach Seminar, November 2008 Examples Mathematical biology: Lotka-Volterra model First numerical methods
More informationPhysics 5153 Classical Mechanics. Canonical Transformations-1
1 Introduction Physics 5153 Classical Mechanics Canonical Transformations The choice of generalized coordinates used to describe a physical system is completely arbitrary, but the Lagrangian is invariant
More informationfor changing independent variables. Most simply for a function f(x) the Legendre transformation f(x) B(s) takes the form B(s) = xs f(x) with s = df
Physics 106a, Caltech 1 November, 2018 Lecture 10: Hamiltonian Mechanics I The Hamiltonian In the Hamiltonian formulation of dynamics each second order ODE given by the Euler- Lagrange equation in terms
More informationOstrogradsky theorem (from 1850)
Ostrogradsky theorem (from 1850) Marián Fecko Department of Theoretical Physics Comenius University Bratislava fecko@fmph.uniba.sk Student Colloqium and School on Mathematical Physics, Stará Lesná, Slovakia,
More informationParallel-in-time integrators for Hamiltonian systems
Parallel-in-time integrators for Hamiltonian systems Claude Le Bris ENPC and INRIA Visiting Professor, The University of Chicago joint work with X. Dai (Paris 6 and Chinese Academy of Sciences), F. Legoll
More informationGeometric Numerical Integration
Geometric Numerical Integration (Ernst Hairer, TU München, winter 2009/10) Development of numerical ordinary differential equations Nonstiff differential equations (since about 1850), see [4, 2, 1] Adams
More informationChap. 3. Controlled Systems, Controllability
Chap. 3. Controlled Systems, Controllability 1. Controllability of Linear Systems 1.1. Kalman s Criterion Consider the linear system ẋ = Ax + Bu where x R n : state vector and u R m : input vector. A :
More informationFINAL EXAM GROUND RULES
PHYSICS 507 Fall 2011 FINAL EXAM Room: ARC-108 Time: Wednesday, December 21, 10am-1pm GROUND RULES There are four problems based on the above-listed material. Closed book Closed notes Partial credit will
More informationPhysical Dynamics (PHY-304)
Physical Dynamics (PHY-304) Gabriele Travaglini March 31, 2012 1 Review of Newtonian Mechanics 1.1 One particle Lectures 1-2. Frame, velocity, acceleration, number of degrees of freedom, generalised coordinates.
More informationHANDOUT #12: THE HAMILTONIAN APPROACH TO MECHANICS
MATHEMATICS 7302 (Analytical Dynamics) YEAR 2016 2017, TERM 2 HANDOUT #12: THE HAMILTONIAN APPROACH TO MECHANICS These notes are intended to be read as a supplement to the handout from Gregory, Classical
More informationDiscrete Dirac Mechanics and Discrete Dirac Geometry
Discrete Dirac Mechanics and Discrete Dirac Geometry Melvin Leok Mathematics, University of California, San Diego Joint work with Anthony Bloch and Tomoki Ohsawa Geometric Numerical Integration Workshop,
More informationCanonical transformations (Lecture 4)
Canonical transformations (Lecture 4) January 26, 2016 61/441 Lecture outline We will introduce and discuss canonical transformations that conserve the Hamiltonian structure of equations of motion. Poisson
More informationTulczyjew triples: from statics to field theory
Tulczyjew triples: from statics to field theory Janusz Grabowski Polish Academy of Sciences 3rd Iberoamerican Meetings on Geometry, Mechanics and Control Salamanca, Sept. 3, 2012 J.Grabowski (IMPAN) Tulczyjew
More informationTHE Hamilton equations of motion constitute a system
Proceedings of the World Congress on Engineering 0 Vol I WCE 0, July 4-6, 0, London, U.K. Systematic Improvement of Splitting Methods for the Hamilton Equations Asif Mushtaq, Anne Kværnø, and Kåre Olaussen
More informationSketchy Notes on Lagrangian and Hamiltonian Mechanics
Sketchy Notes on Lagrangian and Hamiltonian Mechanics Robert Jones Generalized Coordinates Suppose we have some physical system, like a free particle, a pendulum suspended from another pendulum, or a field
More informationElectric and Magnetic Forces in Lagrangian and Hamiltonian Formalism
Electric and Magnetic Forces in Lagrangian and Hamiltonian Formalism Benjamin Hornberger 1/26/1 Phy 55, Classical Electrodynamics, Prof. Goldhaber Lecture notes from Oct. 26, 21 Lecture held by Prof. Weisberger
More informationThe Geometry of Euler s equation. Introduction
The Geometry of Euler s equation Introduction Part 1 Mechanical systems with constraints, symmetries flexible joint fixed length In principle can be dealt with by applying F=ma, but this can become complicated
More informationVariational principles and Hamiltonian Mechanics
A Primer on Geometric Mechanics Variational principles and Hamiltonian Mechanics Alex L. Castro, PUC Rio de Janeiro Henry O. Jacobs, CMS, Caltech Christian Lessig, CMS, Caltech Alex L. Castro (PUC-Rio)
More informationLegendre Transforms, Calculus of Varations, and Mechanics Principles
page 437 Appendix C Legendre Transforms, Calculus of Varations, and Mechanics Principles C.1 Legendre Transforms Legendre transforms map functions in a vector space to functions in the dual space. From
More informationM2A2 Problem Sheet 3 - Hamiltonian Mechanics
MA Problem Sheet 3 - Hamiltonian Mechanics. The particle in a cone. A particle slides under gravity, inside a smooth circular cone with a vertical axis, z = k x + y. Write down its Lagrangian in a) Cartesian,
More information[#1] R 3 bracket for the spherical pendulum
.. Holm Tuesday 11 January 2011 Solutions to MSc Enhanced Coursework for MA16 1 M3/4A16 MSc Enhanced Coursework arryl Holm Solutions Tuesday 11 January 2011 [#1] R 3 bracket for the spherical pendulum
More informationVariational integrators for electric and nonsmooth systems
Variational integrators for electric and nonsmooth systems Sina Ober-Blöbaum Control Group Department of Engineering Science University of Oxford Fellow of Harris Manchester College Summerschool Applied
More informationREVIEW. Hamilton s principle. based on FW-18. Variational statement of mechanics: (for conservative forces) action Equivalent to Newton s laws!
Hamilton s principle Variational statement of mechanics: (for conservative forces) action Equivalent to Newton s laws! based on FW-18 REVIEW the particle takes the path that minimizes the integrated difference
More informationHAMILTON S PRINCIPLE
HAMILTON S PRINCIPLE In our previous derivation of Lagrange s equations we started from the Newtonian vector equations of motion and via D Alembert s Principle changed coordinates to generalised coordinates
More informationBACKGROUND IN SYMPLECTIC GEOMETRY
BACKGROUND IN SYMPLECTIC GEOMETRY NILAY KUMAR Today I want to introduce some of the symplectic structure underlying classical mechanics. The key idea is actually quite old and in its various formulations
More informationCurves in the configuration space Q or in the velocity phase space Ω satisfying the Euler-Lagrange (EL) equations,
Physics 6010, Fall 2010 Hamiltonian Formalism: Hamilton s equations. Conservation laws. Reduction. Poisson Brackets. Relevant Sections in Text: 8.1 8.3, 9.5 The Hamiltonian Formalism We now return to formal
More informationHamiltonian aspects of fluid dynamics
Hamiltonian aspects of fluid dynamics CDS 140b Joris Vankerschaver jv@caltech.edu CDS 01/29/08, 01/31/08 Joris Vankerschaver (CDS) Hamiltonian aspects of fluid dynamics 01/29/08, 01/31/08 1 / 34 Outline
More informationStability of Asynchronous Variational Integrators
Stability of Asynchronous Variational Integrators Abstract The formulation of multiple time step integrators can provide substantial computational savings for mechanical systems with multiple time scales.
More informationPhysical Dynamics (SPA5304) Lecture Plan 2018
Physical Dynamics (SPA5304) Lecture Plan 2018 The numbers on the left margin are approximate lecture numbers. Items in gray are not covered this year 1 Advanced Review of Newtonian Mechanics 1.1 One Particle
More informationPHYS 705: Classical Mechanics. Hamiltonian Formulation & Canonical Transformation
1 PHYS 705: Classical Mechanics Hamiltonian Formulation & Canonical Transformation Legendre Transform Let consider the simple case with ust a real value function: F x F x expresses a relationship between
More informationBalanced models in Geophysical Fluid Dynamics: Hamiltonian formulation, constraints and formal stability
Balanced models in Geophysical Fluid Dynamics: Hamiltonian formulation, constraints and formal stability Onno Bokhove 1 Introduction Most fluid systems, such as the three-dimensional compressible Euler
More informationSynchro-Betatron Motion in Circular Accelerators
Outlines Synchro-Betatron Motion in Circular Accelerators Kevin Li March 30, 2011 Kevin Li Synchro-Betatron Motion 1/ 70 Outline of Part I Outlines Part I: and Model Introduction Part II: The Transverse
More informationM3-4-5 A16 Notes for Geometric Mechanics: Oct Nov 2011
M3-4-5 A16 Notes for Geometric Mechanics: Oct Nov 2011 Text for the course: Professor Darryl D Holm 25 October 2011 Imperial College London d.holm@ic.ac.uk http://www.ma.ic.ac.uk/~dholm/ Geometric Mechanics
More informationHamiltonian Field Theory
Hamiltonian Field Theory August 31, 016 1 Introduction So far we have treated classical field theory using Lagrangian and an action principle for Lagrangian. This approach is called Lagrangian field theory
More informationSome aspects of the Geodesic flow
Some aspects of the Geodesic flow Pablo C. Abstract This is a presentation for Yael s course on Symplectic Geometry. We discuss here the context in which the geodesic flow can be understood using techniques
More informationCourse 241: Advanced Mechanics. Scholarship Questions
Course 41: Advanced echanics Scholarship Questions This covers the first sixteen questions from part I up to the end of rigid bodies, and a selection of the questions from part II. I wouldn t be too bothered
More informationA Very General Hamilton Jacobi Theorem
University of Salerno 9th AIMS ICDSDEA Orlando (Florida) USA, July 1 5, 2012 Introduction: Generalized Hamilton-Jacobi Theorem Hamilton-Jacobi (HJ) Theorem: Consider a Hamiltonian system with Hamiltonian
More information1 Hamiltonian formalism
1 Hamiltonian formalism 1.1 Hamilton s principle of stationary action A dynamical system with a finite number n degrees of freedom can be described by real functions of time q i (t) (i =1, 2,..., n) which,
More informationHamiltonian Mechanics
Chapter 3 Hamiltonian Mechanics 3.1 Convex functions As background to discuss Hamiltonian mechanics we discuss convexity and convex functions. We will also give some applications to thermodynamics. We
More informationUnimodularity and preservation of measures in nonholonomic mechanics
Unimodularity and preservation of measures in nonholonomic mechanics Luis García-Naranjo (joint with Y. Fedorov and J.C. Marrero) Mathematics Department ITAM, Mexico City, MEXICO ẋ = f (x), x M n, f smooth
More informationDUAL HAMILTONIAN STRUCTURES IN AN INTEGRABLE HIERARCHY
DUAL HAMILTONIAN STRUCTURES IN AN INTEGRABLE HIERARCHY RAQIS 16, University of Geneva Somes references and collaborators Based on joint work with J. Avan, A. Doikou and A. Kundu Lagrangian and Hamiltonian
More informationA Hamiltonian Numerical Scheme for Large Scale Geophysical Fluid Systems
A Hamiltonian Numerical Scheme for Large Scale Geophysical Fluid Systems Bob Peeters Joint work with Onno Bokhove & Jason Frank TW, University of Twente, Enschede CWI, Amsterdam PhD-TW colloquium, 9th
More informationComputational Geometric Uncertainty Propagation for Hamiltonian Systems on a Lie Group
Computational Geometric Uncertainty Propagation for Hamiltonian Systems on a Lie Group Melvin Leok Mathematics, University of California, San Diego Foundations of Dynamics Session, CDS@20 Workshop Caltech,
More informationEfficiency of Runge-Kutta Methods in Solving Simple Harmonic Oscillators
MATEMATIKA, 8, Volume 3, Number, c Penerbit UTM Press. All rights reserved Efficiency of Runge-Kutta Methods in Solving Simple Harmonic Oscillators Annie Gorgey and Nor Azian Aini Mat Department of Mathematics,
More informationOn the Hamilton-Jacobi Variational Formulation of the Vlasov Equation
On the Hamilton-Jacobi Variational Formulation of the Vlasov Equation P. J. Morrison epartment of Physics and Institute for Fusion Studies, University of Texas, Austin, Texas 7871-1060, USA. (ated: January
More informationNonlinear Single-Particle Dynamics in High Energy Accelerators
Nonlinear Single-Particle Dynamics in High Energy Accelerators Part 2: Basic tools and concepts Nonlinear Single-Particle Dynamics in High Energy Accelerators This course consists of eight lectures: 1.
More informationNumerical Algorithms as Dynamical Systems
A Study on Numerical Algorithms as Dynamical Systems Moody Chu North Carolina State University What This Study Is About? To recast many numerical algorithms as special dynamical systems, whence to derive
More informationDirac Structures and the Legendre Transformation for Implicit Lagrangian and Hamiltonian Systems
Dirac Structures and the Legendre Transformation for Implicit Lagrangian and Hamiltonian Systems Hiroaki Yoshimura Mechanical Engineering, Waseda University Tokyo, Japan Joint Work with Jerrold E. Marsden
More informationControl, Stabilization and Numerics for Partial Differential Equations
Paris-Sud, Orsay, December 06 Control, Stabilization and Numerics for Partial Differential Equations Enrique Zuazua Universidad Autónoma 28049 Madrid, Spain enrique.zuazua@uam.es http://www.uam.es/enrique.zuazua
More informationIntroduction to numerical methods for Ordinary Differential Equations
Introduction to numerical methods for Ordinary Differential Equations Charles-Edouard Bréhier Abstract The aims of these lecture notes are the following. We introduce Euler numerical schemes, and prove
More informationHamilton-Jacobi theory on Lie algebroids: Applications to nonholonomic mechanics. Manuel de León Institute of Mathematical Sciences CSIC, Spain
Hamilton-Jacobi theory on Lie algebroids: Applications to nonholonomic mechanics Manuel de León Institute of Mathematical Sciences CSIC, Spain joint work with J.C. Marrero (University of La Laguna) D.
More informationTHE LAGRANGIAN AND HAMILTONIAN MECHANICAL SYSTEMS
THE LAGRANGIAN AND HAMILTONIAN MECHANICAL SYSTEMS ALEXANDER TOLISH Abstract. Newton s Laws of Motion, which equate forces with the timerates of change of momenta, are a convenient way to describe mechanical
More informationAlexei F. Cheviakov. University of Saskatchewan, Saskatoon, Canada. INPL seminar June 09, 2011
Direct Method of Construction of Conservation Laws for Nonlinear Differential Equations, its Relation with Noether s Theorem, Applications, and Symbolic Software Alexei F. Cheviakov University of Saskatchewan,
More informationGauge Fixing and Constrained Dynamics in Numerical Relativity
Gauge Fixing and Constrained Dynamics in Numerical Relativity Jon Allen The Dirac formalism for dealing with constraints in a canonical Hamiltonian formulation is reviewed. Gauge freedom is discussed and
More informationThe Particle-Field Hamiltonian
The Particle-Field Hamiltonian For a fundamental understanding of the interaction of a particle with the electromagnetic field we need to know the total energy of the system consisting of particle and
More informationHousekeeping. No announcement HW #7: 3.19
1 Housekeeping No announcement HW #7: 3.19 2 Configuration Space vs. Phase Space 1 A given point in configuration space q q prescribes fully the 1,, n configuration of the system at a given time t. The
More information15. Hamiltonian Mechanics
University of Rhode Island DigitalCommons@URI Classical Dynamics Physics Course Materials 2015 15. Hamiltonian Mechanics Gerhard Müller University of Rhode Island, gmuller@uri.edu Creative Commons License
More informationPhysics 106a, Caltech 13 November, Lecture 13: Action, Hamilton-Jacobi Theory. Action-Angle Variables
Physics 06a, Caltech 3 November, 08 Lecture 3: Action, Hamilton-Jacobi Theory Starred sections are advanced topics for interest and future reference. The unstarred material will not be tested on the final
More informationCocycles and stream functions in quasigeostrophic motion
Journal of Nonlinear Mathematical Physics Volume 15, Number 2 (2008), 140 146 Letter Cocycles and stream functions in quasigeostrophic motion Cornelia VIZMAN West University of Timişoara, Romania E-mail:
More informationChap. 1. Some Differential Geometric Tools
Chap. 1. Some Differential Geometric Tools 1. Manifold, Diffeomorphism 1.1. The Implicit Function Theorem ϕ : U R n R n p (0 p < n), of class C k (k 1) x 0 U such that ϕ(x 0 ) = 0 rank Dϕ(x) = n p x U
More informationM3/4A16. GEOMETRICAL MECHANICS, Part 1
M3/4A16 GEOMETRICAL MECHANICS, Part 1 (2009) Page 1 of 5 UNIVERSITY OF LONDON Course: M3/4A16 Setter: Holm Checker: Gibbons Editor: Chen External: Date: January 27, 2008 BSc and MSci EXAMINATIONS (MATHEMATICS)
More informationClassical mechanics of particles and fields
Classical mechanics of particles and fields D.V. Skryabin Department of Physics, University of Bath PACS numbers: The concise and transparent exposition of many topics covered in this unit can be found
More informationHamiltonian Systems of Negative Curvature are Hyperbolic
Hamiltonian Systems of Negative Curvature are Hyperbolic A. A. Agrachev N. N. Chtcherbakova Abstract The curvature and the reduced curvature are basic differential invariants of the pair: Hamiltonian system,
More informationPart II. Classical Dynamics. Year
Part II Year 28 27 26 25 24 23 22 21 20 2009 2008 2007 2006 2005 28 Paper 1, Section I 8B Derive Hamilton s equations from an action principle. 22 Consider a two-dimensional phase space with the Hamiltonian
More informationOn the construction of discrete gradients
On the construction of discrete gradients Elizabeth L. Mansfield School of Mathematics, Statistics and Actuarial Science, University of Kent, Canterbury, CT2 7NF, U.K. G. Reinout W. Quispel Department
More informationLecture 4: High oscillations
Geometric Numerical Integration TU München Ernst Hairer January February Lecture 4: High oscillations Table of contents A Fermi Pasta Ulam type problem Application of classical integrators 4 3 Exponential
More informationGEOMETRIC QUANTIZATION
GEOMETRIC QUANTIZATION 1. The basic idea The setting of the Hamiltonian version of classical (Newtonian) mechanics is the phase space (position and momentum), which is a symplectic manifold. The typical
More informationDiscretizations of Lagrangian Mechanics
Discretizations of Lagrangian Mechanics George W. Patrick Applied Mathematics and Mathematical Physics Mathematics and Statistics, University of Saskatchewan, Canada August 2007 With Charles Cuell Discretizations
More informationarxiv: v1 [math-ph] 2 Apr 2013
Deviation differential equations. Jacobi fields arxiv:1304.0706v1 [math-ph] 2 Apr 2013 G. SARDANASHVIL Department of Theoretical Physics, Moscow State University, 117234 Moscow, Russia Abstract Given a
More informationEuler-Lagrange Cohomology, Discrete Versions and Applications
November, 2002 Euler-Lagrange Cohomology, Discrete Versions and Applications Han-Ying Guo In collaboration with J.Z. Pan, S.K. Wang, S.Y. Wu, K. Wu and B. Zhou et al. Contents 1 Introduction 2 2 ELCGs
More informationChaotic motion. Phys 750 Lecture 9
Chaotic motion Phys 750 Lecture 9 Finite-difference equations Finite difference equation approximates a differential equation as an iterative map (x n+1,v n+1 )=M[(x n,v n )] Evolution from time t =0to
More informationM. van Berkel DCT
Explicit solution of the ODEs describing the 3 DOF Control Moment Gyroscope M. van Berkel DCT 28.14 Traineeship report Coach(es): Supervisor: dr. N. Sakamoto Prof. dr. H. Nijmeijer Technische Universiteit
More informationarxiv: v1 [physics.plasm-ph] 15 Sep 2013
Symplectic Integration of Magnetic Systems arxiv:309.373v [physics.plasm-ph] 5 Sep 03 Abstract Stephen D. Webb Tech-X Corporation, 56 Arapahoe Ave., Boulder, CO 80303 Dependable numerical results from
More informationSymplectic integration. Yichao Jing
Yichao Jing Hamiltonian & symplecticness Outline Numerical integrator and symplectic integration Application to accelerator beam dynamics Accuracy and integration order Hamiltonian dynamics In accelerator,
More informationGeometric Integrators with Application To Hamiltonian Systems
United Arab Emirates University Scholarworks@UAEU Theses Electronic Theses and Dissertations 9-2015 Geometric Integrators with Application To Hamiltonian Systems Hebatallah Jamil Al Sakaji Follow this
More informationConservation laws for the geodesic equations of the canonical connection on Lie groups in dimensions two and three
Appl Math Inf Sci 7 No 1 311-318 (013) 311 Applied Mathematics & Information Sciences An International Journal Conservation laws for the geodesic equations of the canonical connection on Lie groups in
More informationThe Principle of Least Action
The Principle of Least Action Anders Svensson Abstract In this paper, the principle of least action in classical mechanics is studied. The term is used in several different contexts, mainly for Hamilton
More informationGeometric numerical integration illustrated by the Störmer Verlet method
Acta Numerica (003), pp. 399 450 c Cambridge University Press, 003 DOI: 0.07/S096499000044 Printed in the United Kingdom Geometric numerical integration illustrated by the Störmer Verlet method Ernst Hairer
More informationGeneralized mid-point and trapezoidal rules for Lipschitzian RHS
Generalized mid-point and trapezoidal rules for Lipschitzian RHS Richard Hasenfelder, Andreas Griewank, Tom Streubel Humboldt Universität zu Berlin 7th International Conference on Algorithmic Differentiation
More informationModulation and water waves
Summer School on Water Waves................Newton Institute August 2014 Part 1 Thomas J. Bridges, University of Surrey Modulation of uniform flows and KdV h 0 u 0 00000000000000000000000000000 11111111111111111111111111111
More informationLECTURE 1: LINEAR SYMPLECTIC GEOMETRY
LECTURE 1: LINEAR SYMPLECTIC GEOMETRY Contents 1. Linear symplectic structure 3 2. Distinguished subspaces 5 3. Linear complex structure 7 4. The symplectic group 10 *********************************************************************************
More informationN-Galilean conformal algebras and higher derivatives Lagrangians
arxiv:109.5884v1 [math-ph] 6 Sep 01 N-Galilean conformal algebras and higher derivatives Lagrangians K. Andrzejewski, J. Gonera Department of Theoretical Physics and Computer Science, University of Lódź,
More informationPHYSICS 110A : CLASSICAL MECHANICS
PHYSICS 110A : CLASSICAL MECHANICS 1. Introduction to Dynamics motion of a mechanical system equations of motion : Newton s second law ordinary differential equations (ODEs) dynamical systems simple 2.
More informationValidated Explicit and Implicit Runge-Kutta Methods
Validated Explicit and Implicit Runge-Kutta Methods Alexandre Chapoutot joint work with Julien Alexandre dit Sandretto and Olivier Mullier U2IS, ENSTA ParisTech 8th Small Workshop on Interval Methods,
More informationFrom quantum to classical statistical mechanics. Polyatomic ideal gas.
From quantum to classical statistical mechanics. Polyatomic ideal gas. Peter Košovan peter.kosovan@natur.cuni.cz Dept. of Physical and Macromolecular Chemistry Lecture 5, Statistical Thermodynamics, MC260P105,
More informationChaotic motion. Phys 420/580 Lecture 10
Chaotic motion Phys 420/580 Lecture 10 Finite-difference equations Finite difference equation approximates a differential equation as an iterative map (x n+1,v n+1 )=M[(x n,v n )] Evolution from time t
More informationLectures on Dynamical Systems. Anatoly Neishtadt
Lectures on Dynamical Systems Anatoly Neishtadt Lectures for Mathematics Access Grid Instruction and Collaboration (MAGIC) consortium, Loughborough University, 2007 Part 3 LECTURE 14 NORMAL FORMS Resonances
More informationLecture I: Constrained Hamiltonian systems
Lecture I: Constrained Hamiltonian systems (Courses in canonical gravity) Yaser Tavakoli December 15, 2014 1 Introduction In canonical formulation of general relativity, geometry of space-time is given
More informationRUNGE-KUTTA SCHEMES FOR HAMILTONIAN SYSTEMS
BIT28 (1988),877-8.83 RUNGE-KUTTA SCHEMES FOR HAMILTONIAN SYSTEMS J. M. SANZ-SERNA Departamento de Materndtica Apiicada y Computaci6n, Facuttad de Ciencias, Universidad de Valladolid, Valladolid, Spain
More informationLarge-scale atmospheric circulation, semi-geostrophic motion and Lagrangian particle methods
Large-scale atmospheric circulation, semi-geostrophic motion and Lagrangian particle methods Colin Cotter (Imperial College London) & Sebastian Reich (Universität Potsdam) Outline 1. Hydrostatic and semi-geostrophic
More informationChaos in Hamiltonian systems
Chaos in Hamiltonian systems Teemu Laakso April 26, 2013 Course material: Chapter 7 from Ott 1993/2002, Chaos in Dynamical Systems, Cambridge http://matriisi.ee.tut.fi/courses/mat-35006 Useful reading:
More informationThe dynamical rigid body with memory
The dynamical rigid body with memory Ion Doru Albu, Mihaela Neamţu and Dumitru Opriş Abstract. In the present paper we describe the dynamics of the revised rigid body, the dynamics of the rigid body with
More information