Electronic Journal of Differential Equations, Vol. 2008(2008), No. 71, pp. 1–10. ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu ftp ejde.math.txstate.edu (login: ftp) POWER SERIES SOLUTION FOR THE MODIFIED KDV EQUATION TU NGUYEN Abstract. We use the method developed by Christ [3] to prove local wellposedness of a modified Korteweg de Vries equation in F Ls,p spaces. 1. Introduction The modified Korteweg de Vries (mKdV) equation on a torus T has the form ∂t u + ∂x3 u + u2 ∂x u = 0 u(·, 0) = u0 (1.1) where (x, t) ∈ T × R, u is a real-valued function. If u is a smooth solution of (1.1), 1 then ku(·, t)kL2 (T) = ku0 kL2 (T) for all t; therefore, u e(x, t) = u(x + 2π ku0 k2L2 (T) t, t) is a solution of Z 1 u2 (x, t)dx ∂x u = 0 ∂t u + ∂x3 u + u2 − 2π T (1.2) u(·, 0) = u0 Thus, (1.2) and (1.1) are essentially equivalent. Using Fourier restriction norm method, Bourgain [1] proved that (1.2) is locally well-posed for initial data u0 ∈ H s (T) when s ≥ 1/2, and the solution map is uniformly continuous. In [2], he also showed that the solution map is not C 3 in H s (T) when s < 1/2. Takaoka and Tsutsumi [10] proved local-wellposedness of (1.2) when 1/2 > s > 3/8, and they showed that solution map is not uniformly continuous for this range of s. For (1.1), Kappeler and Topalov [8] used inverse scattering method to show wellposedness when s ≥ 0 and Christ, Colliander and Tao [4] showed that uniformly continuous dependence on the initial data does not hold when s < 1/2. Thus, there is a gap between known local well-posedness results and the space H −1/2 (T) suggested by the standard scaling argument. Recently, Grünrock and Vega [7] showed local well-posedness of the mKdV equation on R with initial data in csr (R) := {f ∈ D0 (R) : kf k cr := kh·is fˆ(·)k r0 < ∞}, H L H s 2000 Mathematics Subject Classification. 35Q53. Key words and phrases. Local well-posedness; KdV equation. c 2008 Texas State University - San Marcos. Submitted April 10, 2008. Published May 13, 2008. 1 2 T. NGUYEN EJDE-2008/71 1 when 2 ≥ r > 1 and s ≥ 12 − 2r . (for r > 43 , this was obtained by Grünrock [5]). This is an extension of the result of Kenig, Ponce and Vega [9] that localcsr scales like H σ wellposedness holds in H s (R) when s ≥ 1/4. Furthermore, as H 1 1 with σ = s + 2 − r , this result covers spaces that have scaling exponent − 21 +. There is also a related recent work of Grünrock and Herr [6] on the derivative nonlinear Schrödinger equation on T. Both [7] and [6] used a version of Bourgain’s method. In this paper, we apply the new method of solution developed by Christ [3] to investigate the local well-posedness of (1.2) with initial data in FLs,p (T) := {f ∈ D0 (T) : kf kF Ls,p := kh·is fˆ(·)klp < ∞}. Let B(0, R) be the ball of radius R centered at 0 in FLs,p (T). Our main result is the following. Theorem 1.1. Suppose s ≥ 1/2, 1 ≤ p < ∞ and p0 (s + 1/4) > 1. Let W be the solution map for smooth initial data of (1.2). Then for any R > 0 there is T > 0 such that the solution map W extends to a uniformly continuous map from B(0, R) to C([0, T ], FLs,p (T)). We note that the FLs,p (T) spaces that are covered by Theorem 1.1 have scaling index 41 +. The restriction s ≥ 1/2 is due to the presence of the derivative in the nonlinear term, and is only used to bound the operator S2 in section 3. The same restriction on s is also required in the work on the derivative nonlinear Schrödinger equation on T by Grünrock and Herr [6]. We believe that the range of p in Theorem 1.1 is not sharp. Concerning (1.1), we have the following result. f be the Corollary 1.2. Suppose s ≥ 1/2, 1 ≤ p < ∞ and p0 (s + 1/4) > 1. Let W solution map for smooth initial data of (1.1). Then for any R > 0 there is T > 0 f extends to a uniformly continuous such that for any c > 0, the solution map W s,p map from B(0, R) ∩ {ϕ : kϕkL2 = c} ⊂ FL (T) to C([0, T ], FLs,p (T)). As in [3], the solution map W obtained in Theorem 1.1 gives a weak solution of (1.2) in the following sense. Let TN be defined by TN u = (χ[−N,N ] u b)∨ . Let R 1 N u := u2 − 2π u2 (x, t)dx ∂x u be the limit in C([0, T ], D0 (T)) of N (TN u) as T N → ∞, provided it exists. Proposition 1.3. Let s and p be as in Theorem 1.1. Let ϕ ∈ FLs,p and u := W ϕ ∈ C([0, T ], FLs,p ). Then N u exists and u satisfies (1.2) in the sense of distribution in (0, T ) × T. To prove these results, we formally expand the solution map into a sum of multilinear operators. These multilinear operators are described in the section 2. Then we will show that if u(·, 0) ∈ FLs,p then the sum of these operators converges in FLs,p for small time t, when s and p satisfy the conditions of Theorem 1.1. Furthermore, this gives a weak solution of (1.2), justifying our formal derivation. EJDE-2008/71 POWER SERIES SOLUTION 3 2. Multilinear operators We rewrite (1.2) as a system of ordinary differential equations of the spatial Fourier series of u (see [10, formula (1.9)], and [1, Lemma 8.16]). dû(n, t) − in3 û(n, t) dt X X = −i û(n1 , t)û(n2 , t)n3 û(n3 , t) + i û(n1 , t)û(−n1 , t)nû(n, t) = n1 +n2 +n3 =n ∗ X −in 3 n n1 (2.1) û(n1 , t)û(n2 , t)û(n3 , t) + inû(n, t)û(−n, t)û(n, t), 1 +n2 +n3 =n where the star means the sum is taken over the triples satisfying nj 6= n, j = 1, 2, 3. We note that these are precisely the triples with σ(n1 , n2 , n3 ) 6= 0. 3 Let a(n, t) = e−in t û(n, t), then an (t) satisfy in da(n, t) =− dt 3 n ∗ X eiσ(n1 ,n2 ,n3 )t a(n1 , t)a(n2 , t)a(n3 , t) 1 +n2 +n3 =n + ina(n, t)a(−n, t)a(n, t), where σ(n1 , n2 , n3 ) = n31 + n32 + n33 − (n1 + n2 + n3 )3 = −3(n1 + n2 )(n2 + n3 )(n3 + n1 ). Or, in integral form, Z ∗ X in t a(n, t) = a(n, 0) − eiσ(n1 ,n2 ,n3 )s a(n1 , s)a(n2 , s)a(n3 , s)ds 3 0 n +n +n =n 1 2 3 (2.2) Z t 2 + in |a(n, s)| a(n, s)ds. 0 If, a is sufficiently nice, say a ∈ C([0, T ], l1 ) (which is the case if u ∈ C([0, T ], H s (T)) for large s) then we can exchange the order of the integration and summation to obtain Z t ∗ X in eiσ(n1 ,n2 ,n3 )s a(n1 , s)a(n2 , s)a(n3 , s)ds a(n, t) = a(n, 0) − 3 n +n +n =n 0 1 2 3 (2.3) Z t + in |a(n, s)|2 a(n, s)ds. 0 Replacing the a(nj , s) in the right hand side by their equations obtained using (2.3), we get Z t ∗ X in a(n1 , 0)a(n2 , 0)a(n3 , 0) eiσ(n1 ,n2 ,n3 )s ds a(n, t) = a(n, 0) − 3 n +n +n =n 0 1 2 3 Z t + in|a(n, 0)|2 a(n, 0) ds + additional terms (2.4) 0 ∗ X n a(n1 , 0)a(n2 , 0)a(n3 , 0) iσ(n1 ,n2 ,n3 )t = a(n, 0) − (e − 1) 3 n +n +n =n σ(n1 , n2 , n3 ) 1 2 3 + int|a(n, 0)|2 a(n, 0) + additional terms 4 T. NGUYEN EJDE-2008/71 The additional terms are those which depend not only on a(·, 0). An example of the additional terms is Z s Z t ∗ ∗ X X 0 nn3 − eiσ(m1 ,m2 ,m3 )s a(n1 , 0)a(n2 , 0) eiσ(n1 ,n2 ,n3 )s 9 n +n +n =n 0 0 m +m +m =n 1 2 3 1 2 3 3 × a(m1 , s0 )a(m2 , s0 )a(m3 , s0 )ds0 ds Then we can again use (2.3) for each appearance of a(m, ·) in the additional terms, and obtain more and more complicated additional terms. We refer to section 2 of [3] for more detailed description of these additional terms. Continuing this process indefinitely, we get a formal expansion of a(n, t) as a sum of multilinear operators of a(·, 0). We will now describe these operators and then show that their sum converges. Again, we refer to section 3 of [3] for more detailed explanations. Each of our multilinear operators will be associated to a tree, which has the property that each of its node has either zero or three children. We will only consider trees with this property. If a node v of T has three children, they will be denoted by v1 , v2 , v3 . We denote by T 0 the set of non-terminal nodes of T , and T ∞ the set of terminal nodes of T . Clearly, if |T | = 3k + 1 then |T 0 | = k and |T ∞ | = 2k + 1. Definition 2.1. Let T be a tree. Then J (T ) is the set of j ∈ ZT such that if v ∈ T 0 then jv = jv1 + jv2 + jv3 , and either jvi 6= jv for all i, or jv1 = −jv2 = jv3 = jv . We will denote by v(T ) be the root of T and j(T ) = j(v(T )). For j ∈ J (T ) and v ∈ T 0 , σ(j, v) := σ(j(v1 ), j(v2 ), j(v3 )). Also define 0 R(T, t) = {s ∈ RT+ : if v < w then 0 ≤ sv ≤ sw ≤ t}. Using the above definitions, we can rewrite (2.4) as X X a(n, t) = a(n, 0) + ωT na(j(v(T )1 ), 0)a(j(v(T )2 ), 0) |T |=4 j∈J (T ),j(T )=n Z × a(j(v(T )3 ), 0) c(j, v(T ), s)ds + additional terms, R(T,t) here c(j, v, s) = eiσ(j,v)s , and ωT is a constant with |ωT | ≤ 1. Continuing this replacement process, it leads to Z X X Y Y a(n, t) = a(n, 0) + ωT ju a(jv , 0) |T |<3k+1 j∈J (T ),j(T )=n u∈T 0 c(j, s)ds R(T,t) v∈T ∞ + additional terms where c(j, s) = Y c(j, v, sv ) v∈T 0 We will show that the series X X a(n, 0) + ωT T Y j∈J (T ),j(T )=n u∈T 0 converges in C([0, T ], lp ) when a(·, 0) ∈ lp . ju Y v∈T ∞ Z a(jv , 0) c(j, s)ds R(T,t) EJDE-2008/71 POWER SERIES SOLUTION 5 3. lp convergence Let T be a tree and j ∈ J (T ). We define Z IT (t, j) = c(j, s)ds, R(T,t) and X ST (t)(av )v∈T ∞ (n) = ωT Y Y ju av (jv )IT (t, j). v∈T ∞ j∈J (T ),j(T )=n u∈T 0 We first give an estimate for IT (t, j) which allows us to bound ST . Lemma 3.1. For 0 ≤ t ≤ 1, |IT (j, t)| ≤ (Ct)|T 0 Y |/2 hσ(j, v)i−1/2 . v∈T 0 Proof. For v ∈ T 0 , define the level of v, denoted l(v), to be the length of the unique path connecting v(T ) and v. Let O be the set of v ∈ T 0 for which l(v) is odd, and E those v for which l(v) is even. First we fix the variables sv with v ∈ E, and take the integration in the variables sv with v ∈ O. For each v ∈ O, the result of the integration is 1 iσ(j,v)sṽ e − eiσ(j,v) max{sv(1) ,sv(2) ,sv(3) } σ(j, v) if σ(j, v) 6= 0, and sṽ − max{sv(1) , sv(2) , sv(3) }. if σ(j, v) = 0. Here ve is the parent of v. Thus, we obtain the factor Y hσ(j, vi−1 v∈O and an integral in sv , v ∈ E where the integrand is bounded by 2|O| . As the domain of integration in sv with v ∈ E has measure less than t|E| , we see that Y 0 |IT (j, t)| ≤ 2|T | t|E| hσ(j, v)i−1 . v∈O By switching the role of O and E, we get 0 |IT (j, t)| ≤ 2|T | t|O| Y hσ(j, v)i−1 . v∈E Combining these two estimates, we obtain the lemma. By Lemma 3.1, |ST (t)(av )v∈T ∞ (n)| ≤ (Ct)|T 0 X |/2 Y hσ(j, u)i−1/2 |ju | Y |av (jv )|. v∈T ∞ j∈J (T ):j(T )=n u∈T 0 Let X SeT (av )v∈T ∞ (n) = Y hσ(j, u)i−1/2 |ju | Y |av (jv )|, v∈T ∞ j∈J (T ):j(T )=n u∈T 0 and e 1 , a2 , a3 )(n) = S(a ∗ X n1 +n2 +n3 =n −1/2 |n|hσ(n1 , n2 , n3 )i 3 Y i=1 |ai (ni )| + |n| 3 Y i=1 |ai (n)|. 6 T. NGUYEN EJDE-2008/71 It is clear that e SeT (av )v∈T ∞ , SeT (av )v∈T ∞ , SeT (av )v∈T ∞ ). SeT (av )v∈T ∞ = S( 1 2 3 1 2 3 where Ti is the subtree of T that contains all nodes u such that u ≤ v(T )i (recall e For this that v(T ) is the root of T ). Hence, to bound ST , it suffices to bound S. purpose, we will use the following simple lemma. Lemma 3.2. Let S be the multilinear operator defined by S(a1 , a2 , a3 )(n) = X m(n1 , n2 , n3 ) n1 +n2 +n3 =n 3 Y aj (nj ), j=1 Let 1 ≤ p ≤ ∞. Then for any pair of indices i 6= j ∈ {1, 2, 3}, kS(a1 , a2 , a3 )klp ≤ sup km(n1 , n2 , n3 )klp0 i,j n 3 Y kak klp . k=1 Proof. By Hölder’s inequality, for any n, |S(a1 , a2 , a3 )(n)| ≤ km(n1 , n2 , n3 )klp0 k i,j 3 Y ≤ sup km(n1 , n2 , n3 )klp0 k n p ak kli,j k=1 i,j 3 Y p ak kli,j k=1 Taking lp -norm in n we obtain the lemma. s Showing that Se is a bounded multilinear map on ls,p := {a : h·i a ∈ lp } is equivalent to showing that S is bounded on lp where S is the operator with kernel hnis |n| m(n1 , n2 , n3 ) = Q3 hσ(n1 , n2 , n3 )i1/2 k=1 hnk is where n1 + n2 + n3 = n. We split S into sum of two operators S1 and S2 where S1 has kernel hnis |n| if n = n1 + n2 + n3 , ni 6= n m1 (n1 , n2 , n3 ) = Q3 s 1/2 k=1 hnk i hn − nk i and S2 has kernel m2 (n1 , n2 , n3 ) = n/hni2s if n1 = −n2 = n3 = n. p Clearly, S2 is bounded on l if and only if s ≥ 1/2. It remains to bound S1 , for which we have the following result. Proposition 3.3. S1 is bounded in lp ×lp ×lp to lp when s ≥ 1/4 and p0 (s+ 41 ) > 1. Proof. In the proof, all the sums are taken over the triples (n1 , n2 , n3 ) that satisfy the additional property that ni 6= n, for all 1 ≤ i ≤ 3. Clearly, we can assume n > 0. Note that if say |n1 | ≥ 5n then as |n2 + n3 | = |n − n1 | ≥ 4n, at least one of n2 and n3 has absolute value bigger than 2n. Also, we cannot have |ni | ≤ n/4 for all i. Thus, up to permutation, there are four cases. (1) |n1 |, |n2 |, |n3 | ∈ [n/4, 5n] (2) |n1 |, |n2 | ∈ [n/4, 5n], |n3 | ≤ n/4 (3) |n1 | ∈ [n/4, 5n], |n2 |, |n3 | ≤ n/4 (4) |n1 |, |n2 | ≥ 2n EJDE-2008/71 POWER SERIES SOLUTION 7 By Lemma 3.2, it suffices to show that in each of these four regions, for some i 6= j p0 the li,j -norm of m is bounded. P Case 1. As 3n = (n − ni ) for some index i, say i = 3, we must have |n − n3 | ∼ n. Since we also have |n1 |, |n2 | & n, |m(n1 , n2 , n3 )| . hni1/2−s . hn3 is |(n − n1 )(n − n2 )|1/2 We will use the inequality | 1 1 1 11 1 1 |≤ . |=| − + n3 (n − n2 ) n1 n3 n − n2 |n1 | |n3 | |n − n2 | 0 0 (1) If 1/4 ≤ s ≤ 1/2: then hn3 ip (1/2−s) . hnip (1/2−s) , so 0 0 X kmkpp0 . l1,2 |n1 |≤5n hnip (1/2−s) |n − n1 |p0 /2 0 p0 /2 |n2 |≤5n 0 X . |n1 |≤5n (hn3 i|n − n2 |) X hnip0 (1/2−s) 1 1 + |n1 |p0 /2 |n − n2 |p0 /2 |n − n1 − n2 |p0 /2 |n2 |≤5n p0 (1−2s) X . hnip (1/2−s) |n − n1 |p0 /2 hn3 ip (1/2−s) X |n1 |≤5n hni An |(n − n1 )n1 |p0 /2 0 . hnip (1−2s) An 0 X 1 1 1 p /2 ( + ) n |n − n1 | |n1 | |n1 |≤5n p0 (1/2−2s) . hni P where A2n . 0 0<j<5n j −p /2 = An . As 1−p0 /2 n An . loghni 1 if p0 < 2 if p0 = 2 if p0 > 2 0 we easily check that hni(1/2−2s)p A2n is bounded by a constant, under our hypothesis on s and p0 . 0 0 (2) If s > 1/2: then hn − n2 ip (s−1/2) . hnip (s−1/2) , so 0 0 kmkpp0 . l1,2 X |n1 |≤5n hnip (1/2−s) |n − n1 |p0 /2 0 X p0 s |n2 |≤5n 0 . X |n1 |≤5n . X |n1 |≤5n hnip (1/2−s) |n − n1 |p0 /2 0 X |n2 |≤5n 0 |n − n1 |p (s−1/2) |n1 |≤5n −p0 /2 . hni (hn3 i|n − n2 |) 1 hnip (s−1/2) 1 + |n1 |p0 s |n − n2 |p0 s |n − n1 − n2 |p0 s Bn |n − n1 |p0 /2 |n1 |p0 s X . Bn hn − n2 ip (s−1/2) Bn2 . 0 1 1 p s + ) n |n − n1 | |n1 | 1 ( 8 T. NGUYEN where Bn = EJDE-2008/71 0 P 0<j<5n j −p s . As 1−p0 s n Bn . loghni 1 if p0 s < 1 if p0 s = 1 if p0 s > 1 0 we easily check that hni−p /2 Bn2 is bounded by a constant, under our hypothesis on s and p0 . Case 2 This case can be treated in exactly the same way as the first case, except when n3 = 0. In the region n3 = 0, 1 X X hnip0 (1/2−s) 0 1 −p0 s kmkpp0 . ≤ hni + l1,3 |n1 (n − n1 )|p0 /2 |n1 |p0 /2 |n − n1 |p0 /2 n n 1 1 −p0 s . hni An . 1 Case 3 As |n1 |, |n − n2 |, |n − n3 | ∼ n, |m(n1 , n2 , n3 )| . 1 . hn2 is hn3 is |n2 + n3 |1/2 Without loss of generality, we assume that |n3 | ≥ |n2 |. (1) If |n2 | < |n3 |/2: X X 0 1 1 kmkpp0 . 0 l2,3 hn2 ip s hn3 ip0 (s+1/2) n/4≥|n3 |>2n2 0≤|n2 |≤n/4 . 1 X 0≤|n2 |≤n/4 hn2 ip0 (2s+1/2)−1 .1 if (s + 1/4)p0 > 1. (2) If |n2 | ≥ |n3 |/2: 0 kmkpp0 . l2,3 X |n3 |≤n/4 . X |n3 |≤n/4 . X |n3 |≤n/4 0 1 hn3 i2p0 s X |n3 |≥n2 ≥|n3 |/2 1 hn3 + n2 ip0 /2 1 −p0 /2+1 } 0 s max{loghn3 i, hn3 i 2p hn3 i loghn3 i + hn3 i2p0 s X |n3 |≤n/4 1 .1 hn3 ip0 (2s+1/2)−1 0 as 2p s ≥ p (s + 1/4) > 1. Case 4 |n1 |, |n2 | > 2n: Note that in this case, |n1 | ∼ |n − n1 | and |n2 | ∼ |n − n3 |. (1) If |n3 |, |n − n3 | ≥ n/2 : we have |m(n1 , n2 , n3 )| . hni1/2 , hn1 is+1/2 hn2 is+1/2 hence 0 0 l1,2 1 X kmkpp0 . hnip /2 |n1 |,|n2 |>2n 0 hn1 hnip /2 . . 1. h2nip0 (2s+1)−2 ip0 (s+1/2) hn p0 (s+1/2) 2i EJDE-2008/71 POWER SERIES SOLUTION 9 (2) If |n3 | < n/2: then |n1 | ∼ |n2 | and |n − n3 | ≥ n/2, so |m(n1 , n2 , n3 )| . hence ns+1/2 , hn1 i2s+1 hn3 is 0 p0 X kmk p0 . Bn l1,3 |n1 |>2n Bn np (s+1/2) . p0 (s+1/2)−1 . 1 hn1 ip0 (2s+1) n (3) If |n − n3 | < n/2: then |n1 | ∼ |n2 | and |n3 | ∼ n. Hence, n . |m(n1 , n2 , n3 )| . 2s+1 hn1 i hn − n3 i1/2 Therefore, 0 0 kmkpp0 . l1,3 X np 0 (2s+1) p hn1 i hn − n3 ip0 /2 X |n1 |≥2n n/2<n3 <3n/2 0 . X |n1 |≥2n An An np . 2p0 s−1 . 1 n hn1 ip0 (2s+1) This concludes the proof of the proposition. Proof of Theorem 1.1. Let u0 ∈ FLs,p and a(n) = u c0 (n). By Proposition 3.3, Y |T 0 | |T 0 |/2 kST ((av )v∈T ∞ )kls,p ≤ C t kav kls,p . v∈T ∞ Hence, ka(n) + X Y j∈J (T ),j(T )=n u∈T 0 T ≤ kakls,p + X ωT X ju Y Z a(jv ) c(j, s)dskls,p R(T,t) v∈T ∞ kST (a, . . . , a)kls,p (3.1) T ≤ ∞ X (Ct)k/2 kak2k+1 ls,p = k=0 ku k s,p √ 0 FL 1 − Ctku0 k2F Ls,p for all t . min{1, ku0 k−4 F Ls,p }. Let T ∼ min{1, ku0 k−4 F Ls,p }, then for t ∈ [0, T ] we can define Z X X Y Y a(n, t) = a(n) + ωT ju a(jv ) T j∈J (T ),j(T )=n u∈T 0 v∈T ∞ c(j, s)ds R(T,t) and the solution map u = W u0 by 3 u b(n, t) = e−in t a(n, t). It follows from (3.1) that W is uniformly continuous. The same argument as that of [3] shows that u is limit of classical solutions. The proof of Proposition 1.2 is basically the same as that of [3, Proposition 1.4], hence we omit it. Acknowledgement. I would like to thank my advisor Prof. Carlos Kenig for suggesting this topic and for his helpful conversations. I would also like to thank Profs. Axel Grünrock and Sebastian Herr for their valuable comments and suggestions. 10 T. NGUYEN EJDE-2008/71 References [1] J. Bourgain; Fourier transform restriction phenomena for certain lattice subsets and applications to nonlinear evolution equations. II. The KdV-equation, Geom. Funct. Anal. 3 (1993), no. 3, 209–262. MR 1215780 (95d:35160b) [2] J. Bourgain; Periodic Korteweg de Vries equation with measures as initial data, Selecta Math. (N.S.) 3 (1997), no. 2, 115–159. MR1466164 (2000i:35173) [3] M. Christ; Power series solution of a nonlinear Schrödinger equation, Mathematical aspects of nonlinear dispersive equations, Ann. of Math. Stud., vol. 163, Princeton Univ. Press, Princeton, NJ, 2007, pp. 131–155. MR 2333210 [4] Michael Christ, James Colliander, and Terrence Tao; Asymptotics, frequency modulation, and low regularity ill-posedness for canonical defocusing equations, Amer. J. Math. 125 (2003), no. 6, 1235–1293. MR 2018661 (2005d:35223) [5] Axel Grünrock; An improved local well-posedness result for the modified KdV equation, Int. Math. Res. Not. (2004), no. 61, 3287–3308. MR 2096258 (2006f:35242) [6] Axel Grünrock and Sebastian Herr; Low regularity local well-posedness of the Derivative Nonlinear Schrödinger Equation with periodic initial data, SIAM J. Math. Anal., to appear. [7] Axel Grünrock and Luis Vega; Local well-posedness for the modified KdV equation in almost cr -spaces, Trans. AMS., to appear. critical H s [8] T. Kappeler and P. Topalov; Global well-posedness of mKdV in L2 (T, R), Comm. Partial Differential Equations 30 (2005), no. 1-3, 435–449. MR 2131061 (2005m:35256) [9] Carlos E. Kenig, Gustavo Ponce, and Luis Vega; Well-posedness and scattering results for the generalized Korteweg-de Vries equation via the contraction principle, Comm. Pure Appl. Math. 46 (1993), no. 4, 527–620. MR 1211741 (94h:35229) [10] Hideo Takaoka and Yoshio Tsutsumi; Well-posedness of the Cauchy problem for the modified KdV equation with periodic boundary condition, Int. Math. Res. Not. (2004), no. 56, 3009– 3040. MR 2097834 (2006e:35295) Tu Nguyen Department of Mathematics, University of Chicago, 5734 S. University Ave., Chicago, IL 60637, USA E-mail address: tu@math.uchicago.edu