# Full text of "The L-equivalent counterpart of the M-III equation"

## See other formats

Preprint CNLP-1994-11 THE L-EQUIVALENT COUNTERPART OF THE M-III EQUATION Q R. Myrzakulov ^] and A. K. Danlybaeva f^] Centre for Nonlinear Problems, PO Box 30, 480035, Alma-Ata-35, Kazakhstan Abstract The connection between differential geometry of curves and the inte- grable (2+1) -dimensional spin system (the M-III equation) is established. Using the proposed geometrical formalism, the L-equivalent counterpart of the M-III equation is found. Preprint CNLP-1994-11. Alma-Ata. 1994 2 E-mail: cnlpmyra@satsun.sci.kz 3 E-mail: cnlpakd@satsun.sci.kz Contents 1 Introduction 2 On the (2+l)-dimensional curve and soliton equations. 3 p.l The M-LXVII equation| .... 3 . . 3 1.2 The M-LXVII equation associated with the M-III equation .3 Particular cases 6 2.3.1 The M-LXVII equation and the M-I equation 6 2.3.2 The M-LXVII equation and the M-II equation 6 3 The M-LXI equation and the M-III equation! 7 |3.1 The M-LXI equation! 7 p.2 The M-LXII equatio^ 7 .3 On the topological invariants .4 The M-III equation |3.5 The L-equivalent of the M-" l [I equation 9 4 The mM-LXI equation and the M-III equation 4.1 The mM-LXI equation 4.2 The mM-LXII equation 4.3 On the topological invariants 4.4 The Lax representation of the mM-LXI equation 4.5 The M-III equation 4.6 The L-equivalent of the M-III equation 5 Geometry of curves and Bilinear representation of the M-III equation 6 The M-III equation as the integrable particular case of the M-0 13 14 14 15 equation 7 The equation for X 8 Conclusion 1 Introduction Consider the Myrzakulov III ( M-III ) equation [1] S t = (S A S y + uS) x + 2l(cl + d)S y - AcvS x + S A W u x = -S ■ (S™ A SA 1 A{2cl + d) 2y x,y W x = JSy (la) (16) (lc) (Id) 1 which was introduced in [1]. Here S = (Si, S2, S3) is the spin vector, S 2 = j3 = ±1, u and v are scalar functions, c, d, I - constants. The M-III equation (1) contains several interesting integrable particular cases: i) c = 0, d = 1, yields the isotropic Myrzakulov I ( M-I ) equation S t = (S A S y + uS) x (2a) U X = S- (S x A Sy). (26) ii) d = yields the Myrzakulov II ( M-II ) equation S t = (S A S y + uS), + 2cl 2 S y - AcvS x (3a) = -S ■ (S x A Sy) (36) hi) c = 0,d = 1, J = diag(0, 0, A), yields the M-I equation with one-ion anisotropy S t = (S A S y + itS) x + 2ZS y + S A W (4a) u x = -S- (S x A S y ) (46) W x = JS y . (4c) iv) The isotropic M-III equation as J = diag(0, 0, 0) S t = (S A S y + uS)^ + 2l(d + d)S y - 4cvS x (5a) u x = -S- (S x A S y ) (5b) v * = A(2d + d )^ l) y (5c) and so on [1]. All of these equations are integrable. For instance, the Lax representation of the isotropic M-III equation (5) has the form [1] 4>x = U'<f> (6a) <h = -2(c\ 2 + d\)&y + V'<l> (66) with where U' = [ic(X 2 - I 2 ) + id(\ - l)]S + °} X l ] sS x , (7a) 2cA -\- d V = [2c(X 2 - I 2 ) + 2d(X - l)]B + X 2 F 2 + XFi + F , (76) F 2 = -4ic 2 VS, Fl = ~ 4icdvs ~ 2cTTd VSS * - MTd S{ss ^ ~ [ss - *l> F = -IFi - l 2 F 2 , B = hy[S, S y ] + 2iuS), S = S-a. 2 In [1] we proposed a new class integrable and nonintegrable spin systems. And we also suggested the geometrical formalism to establish the connection between differential geometry of curves and surfaces and nonlinear evolution equations (NEE), including soliton equations (the A-, B-, C-, D-approaches) (see, also, the refs. [4, 5]). In this paper, using the D-approach we will establish the connection between the differential geometry of curves and the isotropic M- III equation (5). Also we find the L-equivalent (Lakshmanan equivalent [1, 2]) counterpart of this equation. 2 On the (2+l)-dimensional curve and soliton equa- tions. 2.1 The M-LXVII equation Using D-apprach, in [1] we have established the connection between differen- tial geometry of curves and well known soliton equations in 2+1 dimensions. It is remarkable that this approach simultaneously permit determine the L- equivalent counterpart of the under consideration spin systems. Here we will consider a (2+l)-dimensional curves which are given by the M-LXVII equation. The M-LXVII equation reads as [1] (8a) (8b) where we assume that k = k(X, x, y, t), r = r(A, x, y, t), 7 = 7(A, x, y, t), C C(X,x,y,t), D = D(X,x,y,t) and A is some complex parameter, C,D are some matrices. And also let us we suppose that j=i j=i j=i (9a) (9b) C = ]TC^, D = Y,DjXi, e\ = P = ±\, e\ = e\ = \. 3=1 3=1 Here kj = kj(x, y, t), tj = tj(x, y, t), jj = jj(x, y, t), Cj = Cj(x, y, t), Dj = Dj(x,y,t). 2.2 The M-LXVII equation associated with the M-III equation In this subsection we require that the vector ei = S satisfies the M-III equation (5). Then we must put k = -d(q + p), k l = -2c(q+p), kj = 0, j>2 (10a) 3 Here 7o = -d(9-p), 7i = -2c(9-P). 7i = 0, j>2 (106) ro = 0, n = -2d, r 2 = -2c, Tj = 0, j > 3 (10c) Co = 0, Ci = 2d, C 2 = 2c, Cj = j > 3 (lOd) D = A 2 L> 2 + AZ?i + I?o, £>j = j > 3. (lOe) (0 UJ 3 -UJ 2 ^ -Puj 3 wi (11) /?W 2 — Wi y p are some function. So in our case the M-LXVII equation takes the form [1] / -(2cA + d)(g + p) i(2cX + d)(q-p) (3{2cX + d){q + p) -2(cA 2 + dX) v -/3i(2cX + d)(q-p) 2(cA 2 + dA) (ex e 2 ).- ( 61 ^ e 2 V e 3 J ei e 2 e3 ei 2(cA 2 + dX) | e 2 | + (A 2 i9 2 + XD X + D ) e 3 (12a) (126) where the explicit forms of Dj are given in [1] . In terms of matrix this equation we can write in the form where hx = — (2cA + d)(q + p)e 2 + i(2cA + d)(q - p)e 3 e 2x = (3{2cX + d){q + p)ei - 2(cA 2 + dA)e 3 hx = -pi(2cX + d){q- p)h + 2(cA 2 + dA)e 2 e lt = 2(cA 2 + dX)ei y + u; 3 e 2 - oj 2 e 3 e 2t = 2(cA 2 + dX)e 2y - @w 3 e\ + LUie 3 e 3t = 2(cA 2 + dX)e 3y + /3u; 2 ei - wie 2 ei = g 1 (T 3 g, e 2 = g 1 a 2 g, e 3 = g 1 a 1 g. Here <jj are Pauli matrices 1 1 <72 -i 1 •73 1 -1 (13a) (136) (13c) (13d) (13e) (13/) (14) (15) So we have o\ a 2 = ia 3 = —a 2 ai, &i(T 3 = —ia 2 = -a 3 ai, a 3 a 2 = —io\ = -a 2 a 3 (16a) and cr 2 = I = diag(l, 1). (166) Equations (13) we can rewrite in the form [<73,U] = ~(2cA + d)(q + p)a 2 + i(2cA + d)(q - p)a x (17a) 4 Here Hence we get with [(7 2 , U) = (3{2c\ + d)(g + p)a 3 - 2(cA 2 + dA)<n (176) [ai,U] = -[3i(2c\ + d)(q-p)a 3 + 2(c\ 2 + d\)a 2 (17c) [(7 3 , V] = W 3 (7 2 - W 2 <7l (17d) [(7 2 , V] = -/3u; 3 (73 + wio-i (17e) [(71, V] = (3UJ 2 CT3 - Wl(7 2 . (17/) U = g x g- 1 , V = g t g- 1 -2(cX 2 + d\)g x g-\ (18) {7 = i[(cA 2 + dAy 3 + (2cA + d)Q] (19a) y = A 2 5 2 + XB 1 + B (196) J2 d d 2 B 2 = -4ic 2 vcr 3 , Si = -4icdva 3 + 2cQ y a 3 - 8ic 2 vQ, B = -^Bi ~ ~^a B2 ° (20) and Q=(l o)' p = ^- ( 21 ) Thus the matrix-function (7 satisfies the equations 9x = Ug (22a) gt = 2(cA 2 + dX)g y + y 5 (226) The compatibility condition of these equations gives the following M-IIIg equa- tion iqt = q xy - 4ic(vq) x + 2d 2 v q = (23a) — ipt = p X y + 4ic(vp) x + 2d vp = (236) Vx = (pq) y (23c) So we have identified the curve, given by the M-LXVII equation (12) with the M-III equation (5). On the other hand, the compatibilty condition of equa- tions (22) is equivalent to the equation (23). So that we have also established the connection between the curve (the M-LXVII equation) and the equation (23). It means that the M-III equation (5) and the equation (23) are L-equivalent (Lakshmanan equivalent) to each other. 5 2.3 Particular cases 2.3.1 The M-LXVII equation and the M-I equation For the M-I equation (c = 0, d = 1) the associated M-LXVII equation has the form ' —(q + p) i(q~P) P(q + p) -2A | | e 2 | (24a) \-(3i(q-p) 2A I e 2 j V e 3 ) = 2A + D (246) where the explicit form of Do given in [1]. After some algebra we obtain the following L-equivalent of the isotropic M-I equation (2) iqt = q xy + 2vq -ipt = Pxy + 2vp Vx = ipq)y (26a) (266) (26c) It is the Zakharov equation (ZE) [6]. 2.3.2 The M-LXVII equation and the M-II equation Now we put d = 0. Then we get the following version of the M-LXVII equation (27a) /ex x -2cX(q + p) 2ci\(q — p) e 2 ).- 2pc\{q + P) -2cA 2 V e 3 , -2(5ci\{q -P) 2cA 2 / ei ^ (ei \ e 2 = 2cA 2 e 2 V e 3 J t V e 3 / + {\ 2 D 2 + XD Y + D { (276) where the explicit forms of Dj are given in [1]. This M-LXVII equation is associated with the M-II equation (3). Proceeding as above we get the following L-equivalent of the M-II equation iqt = q xy ~ &c(vq) x -ipt = Pxy + 4:ic(vp) a Vx = (Pq)y which is the Strachan equation [7]. (28a) (286) (286) 6 3 The M-LXI equation and the M-III equation 3.1 The M-LXI equation Now, we consider the (2+l)-dimensional curve which is given by the M-LXI equation / e i \ / e i \ (29a) (296) ( 61 ^ [e 3 ) 1- (=) y ( ei \ e2 ve 3 ; where 3.2 The M-LXII equation From (29), we obtain the following M-LXII equation [1] A y -B x + [A, B] = Equation (31a) gives A t -C x + [A,C]=0 B t -C y + [B, C] = k v — m 3x — rm-2 = —m<ix + Tins — km\ = T y — mi x + [3krri2 = 0. The M-LXII equation (31), we can rewrite in form k y - m 3x = /3e 3 • (e 3x A e 3y ) -■mix = (3e 2 ■ (e 2x A e 2y ) (29c) (30a) (306) (30c) (31a) (316) (31c) (32a) (326) (32c) (33a) (336) 7 Ty - mi x = ei • (e lx A e ly ) (33c) Also from (31) we get h - uj 3x = tu>2 = (3e 3 ■ (e 3x A e 3t ) (34a) -u 2x = -tuj 3 + kui = [5e 2 ■ (e 2x A e 2t ) (346) n - LOi x = -f3kuj2 = ei • (ei x A e it ) (34c) and mu - toiy = -[5{m 3 uj 2 - m 2 u 3 ) = ei • {e ly A e w ) (35a) m 2t - u 2y = -miuz + m 3 LOi = [3e 2 ■ (e 2y A e 2t ) (356) m 3t - w 3j/ = -m 2 wi + miio 2 = [3e 3 • (e 3y A e 3t ) (35c) 3.3 On the topological invariants It is interesting to note that the M-LXII equations allows the following integrals of motion [1] K\ = J J K,m 2 dxdy, K 2 = j j Tm 2 dxdy, K 3 = j j (rm 3 — kmi)dxdy (36a) or Ki = J J ei(ei x A e ly )dxdy (376) K 2 = J J e 2 (e 2x Ae 2y )dxdy (376) K 3 = J j e 3 (e 3x A e 3y )dxdy. (37c) So we have the following three topological invariants <2i = 4~ J / ei ' ( eix A e iy)dxdy (38a) Qi = ^ J J e 2 • ( e 2z A e 2y )dxdy (386) Qz = ^ f J e 3 • ( e 3z A e 3y )dxdy (38c) 3.4 The M-III equation Now we wish show how connected the M-LXI (29), M-LXII (31) and M-III (5) equations. Let us we identify S with ei, i.e. ei = S. (39) Then we have mi = u + d x l T y (40a) m 2 = -u x (406) m 3 = 9 X 1 {k y — Tm-i) = d x 1 — (40c) and . uj-v = + tuj 3 ] (41a) ll>2 = —msx — Tm2 + 2l(cl + d)m,2 (416) ^3 = m2x — T~m3 — uk + 2Z(cZ + d)m 3 — Acvk. (41c) 3.5 The L-equivalent of the M-III equation We now introduce the function q according to the following expression k Q 2 (2d + d) expi&licl + djx-d^T}. (42) It is not difficult to verify that the function q satisfy the following nonlinear Schrodinger - type equation iqt = Qxy ~ 4ic(vq) x + 2d 2 v q = (43a) -ipt = Pxy + &c(vp) x + 2d 2 vp = (436) Vx = (Pq)y (43c) It coincide wih the M—III q equation (23). So we have proved that the equations (23) and (5) are L-equivalent to each other. As well known that these equations are G-equivalent to each other [3]. Equation (43) contains two reductions: the Zakharov equation as c = [6] and the Strachan equation as d = [7]. 4 The mM-LXI equation and the M-III equation 4.1 The mM-LXI equation One of the significant model of the (2+l)-dimensional curves is the modified M-LXI (mM-LXI) equation. It reads as [1] (44a) where ( 61 N 'ex \ e2 = A m e2 V e 3 J V e 3 / ( 61 ^ (:) — B m e2 y e2 = C m V e 3 J ( k -13k V fa -r T (446) (44c) (45a) 9 m 3 -TO2 B m = | -/3m 3 mi | ( Lob) (5vri2 —m\ UJ 3 -Ld 2 C m = I U; ~| | ( I •")<") (3ui2 — u\ 4.2 The mM-LXII equation The compatibility condition of the mM-LXI equations (44) gives the modified M-LXII (mM-LXII) equation [1] A my - B mx + [A m , B m ] = (46a) A m t ~ C mx + [A m , C m ] = (466) Bmt-Cmy + [B m ,Cm]=0. (46c) From (46a) we get k y — m 3x + ami — TiTi2 = (47a) <jy — m2x + T1713 — kmi = (476) T y — mi x + (3(km2 — crm 3 ) = 0. (47c) The mM-LXII equation (47a), we can rewrite in form k y - m 3x = /3e 3 • (e 3x A e 3y ) (48a) o-y ~ m 2x = (3e 2 ■ {e 2x A e 2y ) (486) r y - mi x = ei • (ei x A e ly ). (48c) Also from (14) we get h - oj 3x = oloi - toj 2 = +Pe 3 (e 3x A e 3t ) (49a) n - <jJ Vx = (3(kuj2 - cfuj 3 ) = -ei • (eix A e u ) (496) <?t ~ W2x = tw 3 - kwi = [3e 2 ■ (e 2x A e 2t (49c) and ma - uJiy = -f3{m 3 u 2 - m 2 u; 3 ) = e\ ■ (ei y A e u ) (50a) m 2 t - uj 2y = -miu 3 + m 3 u>i = f3e 2 ■ (e 2y A e 2t ) (506) m 3t - uj 3y = -m 2 ui + miw 2 = /3e 3 • (e 2y A e 3t ). (50c) 10 4.3 On the topological invariants It is interesting to note that theM-LXII equations allows the following integrals of motion K\ = j j (nm 2 +am 3 dxdy , K 2 = j j (—Tm 3 +kmi)dxdy , K 3 = J J {rm 2 —am\)dxdy (51a) or Ki = J J ei(ei x A e\ y )dxdy (516) K 2 = J j e 2 (e 2x A e 2y )dxdy (51c) K 3 = J J e 3 (e 3x A e 3y )dxdy (51d) So we have the following three topological invariants Qi = 4" J J e i " ( e i^ A e\y)dxdy (52a) ^ 2 = 4vr / / 62 ' ^ A e2 v"> dxdy ( 52& ) <5s = J J e 3 ■ (e 3s A e 3y )dxdy (52c) 4.4 The Lax representation of the mM-LXI equation To find the Lax representation (LR) of the mM-LXI equation (44) , we rewrite it in the following matrix form e± x = ke 2 — ae 3 (53a) e 2x = —fiei + re 3 (536) e 3x = flaei - re 2 (53c) h y = m 3 e 2 - m 2 e 3 (54a) e 2y = -Pm 3 ei + m\e 3 (546) e 3y = (3m 2 ei - rti\e 2 (54c) iu = oj 3 e 2 - Lo 2 e 3 (55a) e 2 t = -/3uj 3 ei + wie 3 (556) e 3t = f3uj 2 ei - ujie 2 (55c) where ij are given by (14). Equations (53)- (55) we can rewrite in the form (below we put (3 = 1) [03, U] = ko 2 — oo\ (56a) [a 2 ,U] = -pka 3 + Ta 1 (566) [<7i, U] = (3<ja 3 - tu 2 (56c) 11 where Hence we get [03, V] = m 3 (T2 - m 2 ai (57a) [(72, V] = -(3m 3 a 3 - m X (Jx (576) [<7i, V] = /3m 2 cr 3 - mi(72 (57c) [03, W] = W3(J 2 - u; 2 cji (58a) [cr 2 , W] = -/3w 3 (J3 + (586) [(7i, W] = /3LU2U3 - Ui(T2 (58c) U = g x g-\ V = g y g-\ W = g t g-\ (59) u = k{nk T + «,) k -r a ) ^ V = l( M m \ , m3 " im2 ) (606) W = -( M Ul . , W3 ~ iW2 ). (60c) 2i \ 0{w 3 + iwi) -wi J y ' Thus the matrix-function g satisfies the equations g x = Ug, g y = Vg, g t = Wg. (61) This equation is the LR of the mM-LXI equation. Apropos as a = the equation (61) is the LR of the M-LXI equation (29). From the compatibility condition of the equations (61) we get the new form of the mM-LXII equation (46) U y - V x + [U, V}=0 (62a) U t -W x + [17, W] = (626) V t -W y + [V, W]=0 (62c) 4.5 The M-III equation In this subsection, we want establish the connection between the mM-LXI (44), mM-LXII (46) and M-III (5) equations. To this purpose, as above, we identify S with ei, i.e. works the identity (39). Then the identifying variables for the M-III equation (5) are given by mi = u + d~ 1 T y (63a) ^12 = ^(u x + am 3 ) (636) 777.3 = d^i^y + vmi — T77i 2 ) (63c) Ui = - [<j t - lo 2 x + ru) 3 ] (64a) ll> 2 = —m 3x — TIJI2 + ua + 2l(cl + d)?772 — Acva (646) <^3 = m 2x — T m 3 — uk + 2l(d + d)m 3 — 4cv k (64c) and 12 4.6 The L-equivalent of the M-III equation Return to the function q. Let this function has the form k 2 + a 2 q = 2(2cZ + d) expi ^ cl + d ^ x ~ d * ^ ^ 65 ) Then q again satisfies is the equation (23). So, we have again shown that the M-III equation (5) and the equation (23) are L-equivalent to each other. It is remarkable that this result is consistent with the other result namely that these equations are G-equivalent (gauge equivalent) to each other [3]. 5 Geometry of curves and Bilinear representation of the M-III equation In this section we establish self-coordination of the our geometrical formalism that presented above with the other powerful tool of soliton theory - the Hirota's bilinear method. We demonstrate our idea in example the M-III equation (5). For the curve we take the mM-LXI and mM-LXII equations. Usually, for the spin vector S = (S\, S 2 , S3) takes the following transformation S + = S 1 +iS 2 = ^, S 3 = ^^-, A = ff + gg. (66) Also in this section, we assume the (39) is holds. In [4] was shown that for the mM-LXI equation (44) is correct the following representation + 2/5 If -99 ,„„ \ e i = — , ei 3 = ^ (67a) + -P+f J9-I9 ej = 1 — - — , e 23 = 1 ^ (676) + P-9 2 fg + fg f(t7 , 4 = — ^ — . e 33 = ^ — (67c) and , .D x (go f -go f) D x (gof + gof) D x (fof + gog) k = -l 7 , cr = t , T = -l AAA (68a) D y (f°f + 9°9) Dy(g°f + g°f) A D y {g°f-g°f) (686) m\ = -1— m 2 = 7 m 3 = —1 A A A Here = (e^i, ej2, e^), = eji ± ie^- Now we take r = mi = u. (69) Then hence and from (68) we have D x (fof+gog)=0 (70a) So for the M-III equation (5) we have the bilinear representation (66) and (70). This representation allows construct the bilinear form of the M-III equation which is left in future. 13 6 The M-III equation as the integrable particular case of the M-0 equation Consider the (2+l)-dimensional M-0 equation [1] S* = ai 2 e 2 + ai 3 e 3 , 8^ = 61262 + 61363, S y = ci 2 e 2 + ci 3 e 3 (71) where e2 = ~^ S x ~ ~X S y ' es = ~~X Sx + "If Sj/ ' A = bl2 ° 13 ~ b ^ci2- (72) All known spin systems (integrable and nonintegrable) in 2+1 dimensions are the particular reductions of the M-0 equation (71). In particular, the M-III equation (5) is the integrable reduction of equation (71). In this case, we have a i2 = v 3 , ai 3 = -co 2 , 612 = k, 613 = -a, c i2 = m 3 , ci 3 = -m 2 . (73) Sometimes we use the following form of the M-0 equation [1] S t = d 2 S x + d 3 S y (74) with , 012C13 - ai 3 ci2 , a u b 13 - a 13 b 12 d2 = X » d 3 = X • ( 75 ) A A 7 The equation for A Let us consider the M-LXVII equation in he form (12) for the case q = p = 0. We have (76a) ° ^ -2(cA 2 + dA) 2(cA 2 + dX) / : ei 2(cA 2 + dX) I e 2 e3 Hence we get (766) (77) A t = 2(cA 2 + dX)X y So for the M-III equation (5) the spectral parameter satisfies the equation (77), i.e. in this case we have a nonisospectral problem. From (77) we obtain 1) for the M-I equation (2) 2) for the M-II equation (3) Aj — 2XX y . X t = 2cX 2 X v Now consider the general form of such equations Xt = kX n X y , k = const. (78) (79) (80) 14 This equation has the following solution where a(c) is real (complex) constant. 8 Conclusion In this paper, we have used a geometrical approach pioneered by Lakshmanan to analyze the connection between differential geometry of curves and spin sys- tems to establish such connection with the integrable (2+1)- dimensional spin system- called M-III equation. Simultaneously our approach permit construct the corresponding L-equivalent of the given spin system and Lax representation of it. Some other consequences of this geometrical formalism are presented. References [1] Myrzakulov R 1987 On some integrable and nonintegrable soliton equations of magnets Preprint (Alma-Ata: HEPI) [2] Lakshmanan M 1977 Phys. Lett. 61A 53 [3] Nugmanova G N 1992 The Myrzakulov equations: the gauge equivalent counterparts and soliton solutions (Alma-Ata: KSU) [4] Myrzakulov R 1994 Soliton equations in 2+1 dimensions and differential geometry of curves/surfaces Preprint CNLP-1994-02 (Alma-Ata: CNLP) [5] Myrzakulov R and Syzdykova R N 1994 On the L-equivalence bettween the Ishimori equation and the Davey-Stewartson equation Preprint CNLP- 1994-10 (Alma-Ata: CNLP) [6] Zakharov V E in Solitons ed. R. K. Bullough and P. J. Caudrey ( Springer, Berlin, 1980) [7] Strachan I A B 1993 J. Math. Phys. 34 243 15