Symbolic Computation of Conservation Laws of Nonlinear Partial Differential Equations in Multiple Space Dimensions Willy Hereman Department of Mathematical and Computer Sciences Colorado School of Mines Golden, Colorado Symmetry Plus Integrability 2010 South Padre Island, Texas Monday, June 14, 2010, 11:30a.m. Acknowledgements Douglas Poole (Ph.D. Dissertation, CSM) Mark Hickman (Univ. of Canterbury, New Zealand) Bernard Deconinck (Univ. of Washington, Seattle) Several undergraduate and graduate students Research supported in part by NSF under Grant No. CCF-0830783 This presentation was made in TeXpower Outline • Why compute conservation laws? • The Korteweg-de Vries and Zakharov-Kuznetsov equations • Demonstration of ConservationLawsMD.m • Algorithm for computing conservation laws • Tools from the calculus of variations • Applicaton to the Zakharov-Kuznetsov equation • Conclusions and future work • Additional examples I Manakov-Santini system I Camassa-Holm equation in (2+1) dimensions I Khoklov-Zabolotskaya equation I Shallow water wave model for atmosphere I Kadomtsev-Petviashvili equation I Potential Kadomtsev-Petviashvili equation I Generalized Zakharov-Kuznetsov equation Conservation Laws for Nonlinear PDEs • System of evolution equations of order M ut = F(u(M ) (x)) with u = (u, v, w, . . .) and x = (x, y, z). • Conservation law in (1+1)-dimensions Dt ρ + D x J = 0 evaluated on the PDE. Conserved density ρ and flux J. • Conservation law in (2+1)-dimensions D t ρ + ∇ · J = D t ρ + D x J1 + D y J2 = 0 evaluated on the PDE. Conserved density ρ and flux J = (J1 , J2 ). • Conservation law in (3+1)-dimensions Dt ρ + ∇ · J = Dt ρ + Dx J1 + Dy J2 + Dz J3 = 0 evaluated on the PDE. Conserved density ρ and flux J = (J1 , J2 , J3 ). Reasons for Computing Conservation Laws • Conservation of physical quantities (linear momentum, mass, energy, electric charge, ... ) • Testing of complete integrability and application of Inverse Scattering Transform • Testing of numerical integrators • Study of quantitative and qualitative properties of PDEs (Hamiltonian structure, recursion operators, . . . ) • Verify the closure of a model Examples of PDEs with Conservation Laws • Korteweg-de Vries (KdV) equation models shallow water waves, ion-acoustic waves in plasmas, etc. ∂u ∂u ∂3u +u + =0 3 ∂t ∂x ∂x Diederik Korteweg or ut + uux + uxxx = 0 Gustav de Vries • First six (of infinitely many) conservation laws: Dt u + Dx 12 u2 + uxx = 0 Dt u2 + Dx 32 u3 − ux 2 + 2uuxx = 0 Dt u3 − 3u2x + Dx 43 u4 − 6uu2x + 3u2 uxx + 3u2xx − 6ux uxxx = 0 Dt u4 − 12uu2x + 36 2 u 5 xx − 18uu2x + 4u3 uxx 2 36 2 72 +12u2x uxx + 96 uu −24uu u − u + u u =0 x xxx xx 5 5 xxx 5 xx 4x + Dx 4 5 u 5 2 Dt u5 − 30 u2 u2x + 36 uu2xx − 108 u xxx 7 + Dx 56 u6 − 40u3 u2x − . . . − 216 uxxx u5x = 0 7 Dt u6 − 60 u3 u2x − 30 u4x + 108 u2 u2xx 3 2 648 216 2 − + u uu u4x + 720 xx xxx 7 7 7 + Dx 76 u7 − 75u4 u2x − . . . + 432 u4x u6x = 0 7 • Third conservation law: Gerald Whitham, 1965 • Fourth and fifth: Norman Zabusky, 1965-66 Seventh (sixth thru tenth): Robert Miura, 1966 • Robert Miura • Conservation law explicitly dependent on t and x: Dt tu2 − 2xu +Dx 23 tu3 − xu2 + 2ux − tu2x + 2tuuxx − 2xuxx = 0 First eleven densities: Control Data Computer CDC-6600 computer (2.2 seconds) −→ large integers problem! Control Data CDC-6600 • Zakharov-Kuznetsov (ZK) equation models ion-sound solitons in a low pressure uniform magnetized plasma ut + αuux + β(uxx + uyy )x = 0 • Conservation laws: Dt u + Dx α2 u2 + βuxx + Dy βuxy = 0 3 2 2 Dt u2 + Dx 2α u − β(u − u x y ) + 2βu(uxx + uyy ) 3 − Dy 2βux uy = 0 More conservation laws (ZK equation): 2 2 2 2 2 3α 4 + u ) + D + u Dt u3 − 3β (u u + 3βu u − 6βu(u x xx x y x y) α 4 2 2 + 3βα (u2xx − u2yy ) − 6βα (ux (uxxx + uxyy ) + uy (uxxy + uyyy )) 2 + Dy 3βu2 uxy + 6βα uxy (uxx + uyy ) = 0 3 2 2 Dt tu2 − α2 xu + Dx t( 2α u − β(u − u x y ) + 2βu(uxx + uyy )) 3 2β 1 − x(u2 + 2β u ) + u − D 2β(tu u + xuxy ) = 0 y x y α xx α x α Methods for Computing Conservation Laws • Use the Lax pair L and A, satifying [L, A] = 0. If L = Dx + U, A = Dt + V then Vx − Ut + [U, V ] = 0. L̂ = T LT −1 gives the densities, Â = T AT −1 gives the fluxes. • Use Noether’s theorem (Lagrangian formulation) to generate conservation laws from symmetries (Ovsiannikov, Olver, Rosenhaus, and many others). • Integrating factor methods (Anderson, Bluman, Anco, Cheviakov, Wolf, etc.) require solving ODEs (or PDEs). Proposed Algorithmic Method • Density is linear combination of scaling invariant terms (in the jet space) with undetermined coefficients. • Compute Dt ρ with total derivative operator. • Use variational derivative (Euler operator) to express exactness. • Solve a (parametrized) linear system to find the undetermined coefficients. • Use the homotopy operator to compute the flux (invert Dx or Div). • Work with linearly independent pieces in finite dimensional spaces. • Use linear algebra, calculus, and variational calculus (algorithmic). • Implement the algorithm in Mathematica. Software Demonstration Notation – Computations on the Jet Space • Independent variables x = (x, y, z) • Dependent variables u = (u(1) , u(2) , . . . , u(j) , . . . , u(N ) ) In examples: u = (u, v, θ, h, . . .) • ∂k u Partial derivatives ukx = ∂xk , ∂5u Examples: uxxxxx = u5x = ∂x5 ∂6u uxx yyyy = u2x4y = ∂x2 y4 • u(M ) represents all components of u and all its partial derivatives up to order M. • Differential functions Example: f = uvvx + x2 u3x vx + ux vxx ukx ly = ∂ k+l u , ∂xk y l etc. Tools from the Calculus of Variations • Definition: A differential function f is a exact iff f = DivF. Special case (1D): f = Dx F. • Question: How can one test that f = DivF ? • Theorem (exactness test): f = Div F iff Lu(j) (x) f ≡ 0, j = 1, 2, . . . , N. N is the number of dependent variables. The Euler operator annihilates divergences • Euler operator in 1D (variable u(x)): M X k ∂ Lu(x) = (−Dx ) ∂ukx k=0 ∂ ∂ ∂ 2 ∂ 3 = − Dx + Dx − Dx + ··· ∂u ∂ux ∂uxx ∂uxxx • Euler operator in 2D (variable u(x, y)): Lu(x,y) = My Mx X X k=0 ∂ (−Dx ) (−Dy ) ∂u kx `y `=0 k ` ∂ ∂ ∂ = − Dx − Dy ∂u ∂ux ∂uy ∂ ∂ 2 ∂ 2 ∂ 3 + Dx +Dx Dy +Dy −Dx ··· ∂uxx ∂uxy ∂uyy ∂uxxx • Question: How can one compute F = Div−1 f ? • Theorem (integration by parts): • In 1D: If f is exact then F = D−1 x f = Z f dx = Hu(x) f • In 2D: If f is a divergence then (x) (y) F = Div−1 f = (Hu(x,y) f, Hu(x,y) f ) The homotopy operator inverts total derivatives and divergences! • Homotopy Operator in 1D (variable x): Z Hu(x) f = 0 N 1X dλ (Iu(j) f )[λu] λ j=1 with integrand (j) Mx Iu(j) f = X k−1 X k=1 (j) uix k−(i+1) (−Dx ) i=0 ∂f (j) ∂ukx (Iu(j) f )[λu] means that in Iu(j) f one replaces u → λu, ux → λux , etc. More general: u → λ(u − u0 ) + u0 ux → λ(ux − ux 0 ) + ux 0 etc. Application to Zakharov-Kuznetsov Equation ut + αuux + β(uxx + uyy )x = 0 • Step 1: Compute the dilation invariance ZK equation is invariant under scaling symmetry t x y 2 (t, x, y, u) → ( 3 , , , λ u)= (t̃, x̃, ỹ, ũ) λ λ λ λ is an arbitrary parameter. • Assign weights to each variable W (u) = 2, W (Dt ) = 3, W (Dx ) = 1, W (Dy ) = 1. • Rank of a monomial is the sum of the weights of the variables. Example: Rank(αuux ) = 2W (u) + W (Dx ) = 5. • Key observation: A conservation law is invariant under the scaling symmetry of the PDE W (u) = 2, W (Dt ) = 3, W (Dx ) = 1, W (Dy ) = 1. For example, 2 2 2 2 2 3α 4 +u ) + D + u Dt u3 − 3β (u u +3βu u −6βu(u x 4 xx x y x y) α 2 2 + 3βα (u2xx −u2xy ) − 6βα (ux (uxxx +uxyy ) + uy (uxxy +uyyy )) 2 6β 2 + Dy 3βu uxy + α uxy (uxx + uyy ) = 0 Rank(ρ) = 6, Rank(J) = 8. Rank (conservation law) = 9. Rank of the density needs to be selected! • Step 2: Construct the candidate density For example, construct a density of rank 6. Make a list of all terms with rank 6: {u3 , u2x , uuxx , u2y , uuyy , ux uy , uuxy , u4x , u3xy , u2x2y , ux3y , u4y } Remove divergences and divergence-equivalent terms. Candidate density of rank 6: ρ = c1 u3 + c2 u2x + c3 u2y + c4 ux uy • Step 3: Compute the undetermined coefficients Compute ∂ρ Dt ρ = + ρ0 (u)[ut ] ∂t My Mx X X ∂ρ ∂ρ = + Dkx D`y ut ∂t ∂ukx `y k=0 `=0 = 3c1 u2 I + 2c2 ux Dx + 2c3 uy Dy + c4 (uy Dx + ux Dy ) ut Substitute ut = − αuux + β(uxx + uyy )x . E = −Dt ρ = 3c1 u2 (αuux + β(uxx + uxy )x ) + 2c2 ux (αuux + β(uxx + uyy )x )x + 2c3 uy (αuux + β(uxx + uyy )x )y + c4 (uy (αuux + β(uxx + uyy )x )x + ux (αuux + β(uxx + uyy )x )y ) Apply the Euler operator (variational derivative) My Mx X X k ` ∂E Lu(x,y) E = (−Dx ) (−Dy ) ∂ukx `y k=0 `=0 = −2 (3c1 β + c3 α)ux uyy + 2(3c1 β + c3 α)uy uxy +2c4 αux uxy +c4 αuy uxx +3(3c1 β +c2 α)ux uxx ≡0 Solve a parameterized linear system for the ci : 3c1 β + c3 α = 0, c4 α = 0, 3c1 β + c2 α = 0 Solution: 3β , c = − , c4 = 0 c1 = 1, c2 = − 3β 3 α α Substitute the solution into the candidate density ρ = c1 u3 + c2 u2x + c3 u2y + c4 ux uy Final density of rank 6: 3β 2 ρ=u − (ux + u2y ) α 3 • Step 4: Compute the flux Use the homotopy operator to invert Div: (x) (y) J = Div−1 E = Hu(x,y) E, Hu(x,y) E Z 1 where dλ (x) (x) (Iu E)[λu] Hu(x,y) E = λ 0 with Iu(x) E = My k−1 ` Mx X X XX k=1 `=0 (−Dx ) i+j k+`−i−j−1 k−i−1 uix jy i k+` i=0 j=0 k k−i−1 `−j ∂E −Dy ∂ukx `y (y) (y) Similar formulas for Hu(x,y) E and Iu E. Let A = αuux + β(uxxx + uxyy ) so that E = 3u2 A − 6β u A α x x − 6β u A α y y Then, (x) (y) J = Hu(x,y) E, Hu(x,y) E = 43 αu4 + βu2 (3uxx + 2uyy ) − βu(6u2x + 2u2y ) + − + − 3β 2 β2 7 u(u + u ) − u ( u + 6uxxx ) 2x2y 4y 4α α x 2 xyy β2 β2 2 3 5 2 3 2 u (4u + u ) + (3u + u + u ) y xxy yyy xx α 2 α 2 xy 4 yy 5β 2 2 u u , βu uxy − 4βuux uy xx yy 4α 3β 2 β2 u(ux3y + u3xy ) − 4α ux (13uxxy + 3uyyy ) 4α − 4α uy (uxxx + 3uxyy ) + 4α uxy (uxx + uyy ) 5β 2 9β 2 However, Div−1 E is not unique. Indeed, J = J̃ + K, where K = (Dy θ, −Dx θ) is a curl term. For example, 2 β θ = 2βu2 uy + 4α 3u(uxxy +uyyy )+10ux uxy +5uy (3uyy +uxx ) Shorter flux: J̃ = J − K 4 2 2 2 u + 3βu u − 6βu(u + u = 3α xx x y) + 4 − 6β 2 (ux (uxxx α 3βu2 uxy 3β 2 α u2xx − u2yy + uxyy ) + uy (uxxy + uyyy )), 2 + 6βα uxy (uxx + uyy ) Additional Examples • Manakov-Santini system utx + uyy + (uux )x + vx uxy − uxx vy = 0 vtx + vyy + uvxx + vx vxy − vy vxx = 0 • Conservation laws for Manakov-Santini system: Dt f ux vx + Dx f (uux vx − ux vx vy − uy vy ) − f 0 y(ut + uux − ux vy ) + Dy f (ux vy + uy vx + ux vx2 ) + f 0 (u − yuy − yux vx ) = 0 where f = f (t) is arbitrary. Conservation laws – continued: Dt f (2u + vx2 − yux vx ) + Dx f (u2 + uvx2 + uy v − vy2 − vx2 vy − y(uux vx − ux vx vy − uy vy )) − f 0 y(vt + uvx − vx vy ) − (2f x − f 0 )y 2 (ut + uux − ux vy ) − Dy f (ux v − 2vx vy − vx3 + y(ux vx2 + ux vy + uy vx )) −f 0 (v − y(2u + vy + vx2 )) + (2f x − f 0 y 2 )(ux vx + uy ) = 0 where f = f (t) is arbitrary. There are three additional conservation laws. • (2+1)-dimensional Camassa-Holm equation (αut + κux − utxx + 3βuux − 2ux uxx − uuxxx )x + uyy = 0 Interchange t with y (αuy + κux − uxxy + 3βuux − 2ux uxx − uuxxx )x + utt = 0 Set v = ut to get ut = v vt = −αuxy − κuxx + u3xy − 3βu2x − 3βuuxx + 2u2xx + 3ux uxxx + uu4x • Conservation laws for the Camassa-Holm equation Dt f u + Dx 2 1 3 f ( βu α 2 + κu − 12 u2x − uuxx − utx ) − ( α1 f x − 12 f 0 y 2 )(αut + κux + 3βuux − 2ux uxx − uuxxx − utxx ) − Dy uy ( α1 f x − 12 f 0 y 2 ) + f 0 yu = 0 Dt f yu + Dx α1 f y( 32 βu2 + κu − 21 u2x − uuxx − utx ) − y( α1 f x − 16 f 0 y 2 )(αut + κux + 3βuux − 2ux uxx − uuxxx − utxx ) − Dy yuy ( α1 f x − 61 f 0 y 2 ) − u( α1 f x − 12 f 0 y 2 ) = 0 where f = f (t) is an arbitrary function. • Khoklov-Zabolotskaya equation describes e.g. sound waves in nonlinear media (ut − uux )x − uyy − uzz = 0 Conservation law: Dt f u − Dx 12 f u2 + (f x + g)(ut − uux ) − Dy (fy x + gy )u − (f x + g)uy − Dz (fz x + gz )u − (f x + g)uz = 0 under the constraints ∆f = 0 and ∆g = ft where f = f (t, y, z) and g = g(t, y, z). • Shallow water wave model (atmosphere) ut + (u·∇)u + 2 Ω × u + ∇(θh) − 12 h∇θ = 0 θt + u·(∇θ) = 0 ht + ∇·(uh) = 0 where u(x, y, t), θ(x, y, t) and h(x, y, t). • In components: ut + uux + vuy − 2 Ωv + 12 hθx + θhx = 0 vt + uvx + vvy + 2 Ωu + 21 hθy + θhy = 0 θt + uθx + vθy = 0 ht + hux + uhx + hvy + vhy = 0 • First few conservation laws of SWW model: u (1) ρ(1) = h J =h v u (2) ρ(2) = h θ J = hθ v u 2 (3) 2 ρ(3) = h θ J = hθ v 2 + v 2 + 2hθ) u (u ρ(4) = h (u2 + v 2 + hθ) J(4) = h v (v 2 + u2 + 2hθ) ρ(5) = θ (2Ω + vx − uy ) 4Ωu − 2uuy + 2uvx − hθy 1 (5) J =2θ 4Ωv + 2vvx − 2vuy + hθx • More general conservation laws for SWW model: Dt f (θ)h + Dx f (θ)hu + Dy f (θ)hv = 0 Dt g(θ)(2Ω + vx − ux ) + Dx 21 g(θ)(4Ωu − 2uuy + 2uvx − hθy ) + Dy 12 g(θ)(4Ωv − 2uy v + 2vvx + hθx ) = 0 for any functions f (θ) and g(θ). • Kadomtsev-Petviashvili (KP) equation (ut + αuux + uxxx )x + σ 2 uyy = 0 parameter α ∈ IR and σ 2 = ±1. Equation be written as a conservation law Dt (ux ) + Dx (αuux + uxxx ) + Dy (σ 2 uy ) = 0. Exchange y and t and set ut = v ut = v vt 1 = − 2 (uxy + αu2x + αuuxx + uxxxx ) σ • Examples of conservation laws for KP equation (explicitly dependent on t, x, and y) Dt xux +Dx 3u2 −uxx −6xuux +xuxxx +Dy αxuy = 0 Dt yux +Dx y(αuux + uxxx ) +Dy σ 2 (yuy − u) = 0 √ √ 2y2 2y2 √ σ σ Dt tu + Dx 12 α tu2 + tuxx + √ ut + √ uxxx 4 t 4 t √ √ √ ασ 2 y 2 + √ uux − x tut − αx tuux − x tuxxx 4 t √ yu y 2 uy +Dy − √ + √ + x tuy = 0 2 t 4 t • More general conservation laws for KP equation: Dt f (t)u + Dx f (t)( 12 αu2 + uxx ) +( 21 σ 2 f 0 (t)y 2 − f (t)x)(ut + αuux + u3x ) +Dy uy ( 21 f 0 (t)y 2 − σ 2 f (t)x) − f 0 (t)yu = 0 Dt f (t)yu + Dx f (t)y( 21 αu2 + uxx ) +( 61 σ 2 f 0 (t)y 3 − f (t)xy)(ut + αuux + u3x ) +Dy uy ( 61 f 0 (t)y 3 − σ 2 f (t)xy) −u( 21 f 0 (t)y 2 − σ 2 f (t)x) = 0 where f (t) is arbitrary function. • Potential KP equation Replace u by ux and integrate with respect to x. uxt + αux uxx + uxxxx + σ 2 uyy = 0 • Examples of conservation laws (not explicitly dependent on x, y, t): Dt ux + Dx 12 αu2x + uxxx + Dy σ 2 uy = 0 Dt u2x + Dx 23 αu3x − u2xx + 2ux uxxx − σ 2 uyy +Dy 2σ 2 ux uy = 0 Conservation laws for pKP equation – continued: Dt ux uy + Dx αu2x uy + ut uy + 2uxxx uy − 2uxx uxy +Dy σ 2 u2y − 31 u3x − ut ux + u2xx = 0 Dt 2αuux uxx + 3uu4x − 3σ 2 u2y + Dx 2αut u2x + 3u2t − 2αuux utx − 3utx uxx + 3ut uxxx + 3ux ut xx − 3uut xxx +Dy 6σ 2 ut uy = 0 Various generalizations exist. • Generalized Zakharov-Kuznetsov equation ut + αun ux + β(uxx + uyy )x = 0 where n is rational, n 6= 0. Conservation laws: α Dt u + Dx n+1 un+1 + βuxx + Dy βuxy = 0 2α n+2 Dt u2 + Dx n+2 u − β(u2x − u2y ) + 2βu(uxx + uyy ) − Dy 2βux uy = 0 • Third conservation law for gZK equation: (n+1)(n+2)β n+2 2 2 − Dt u (u + u x y) 2α (n+2)α 2(n+1) + (n + 2)βun+1 uxx + Dx 2(n+1) u (u2x + u2y ) (n+1)(n+2)β 2 2 (u xx 2α − u2yy ) 2 − (n+1)(n+2)β (ux (uxxx + uxyy ) + uy (uxxy + uyyy )) α 2 + Dy (n + 2)βun+1 uxy + (n+1)(n+2)β uxy (uxx + uyy ) = 0. α − (n + 1)(n + 2)βu n + Conclusions and Future Work • • The power of Euler and homotopy operators: I Testing exactness I −1 Integration by parts: D−1 and Div x Integration of non-exact expressions Example: f = ux v + uvx + u2 uxx R R 2 f dx = uv + u uxx dx • Use other homotopy formulas (moving terms amongst the components of the flux; prevent curl terms) • Broader class of PDEs (beyond evolution type) Example: short pulse equation (nonlinear optics) uxt = u + (u3 )xx = u + 6uu2x + 3u2 uxx with non-polynomial conservation law p p Dt 1 + 6u2x − Dx 3u2 1 + 6u2x = 0 • Continue the implementation in Mathematica • Software: http://inside.mines.edu/∼whereman Thank You Gelukkige Verjaardag Yuji