MATH 220: DISTRIBUTIONS AND WEAK DERIVATIVES

ANDRAS VASY

Suppose V is a vector space over F = R or F = C. The algebraic dual of V is the vector space L(V, F) consisting of linear functionals from V to F. That is elements of f ∈ L(V, F) are linear maps f : V → F satisfying f (v + w) = f (v) + f (w), f (cv) = cf (v), v, w ∈ V, c ∈ F. When V is infinite dimensional, we need additional information, namely continuity. So if V is a topological space with the compatible with the vector space structure, i.e. if V is a topological vector space, we define the dual space V ∗ as the space of continuous linear maps f : V → F. For us, V is the class of ‘very nice objects’, and V ∗ will be the class of ‘bad objects’. Of course, normally there is no way of comparing elements of V with those of V ∗ , so we will also need an injection ι:V →V∗ so that elements of V can be regarded as elements of V ∗ (by identifying v ∈ V with ι(V ). As we want to differentiate functions, as much as we desire, V will consist of infinitely differentiable functions. As we need to control behavior at infinity to integrate, the elements of V will be compactly supported. So we define ∞ V = Cc (Rn ) to be the space of infinitely continuously differentiable functions of compact support. Here recall that the support supp φ of continuous function φ is ∞ the closure of the set where φ = 0; so φ ∈ Cc (Rn ) means that there is a compact n (i.e. closed and bounded) subset K of R such that φ ≡ 0 outside K. ∞ As Cc (Rn ) is infinite dimensional, we also need to put a topology on this. Technically this means that we should define what open sets are. Rather than doing this (to avoid complexity) we define what convergence of a sequence φj of functions ∞ in Cc (Rn ) means.
∞ ∞ Definition 1. Suppose {φj }j=1,2,... is a sequence in Cc (Rn ), and φ ∈ Cc (Rn ). We say that limj→∞ φj = φ if (i) there is a compact set K such that φj ≡ 0 outside K for all j, (ii) and all derivatives of φj converge uniformly to φ, i.e. for all multiindices α, supRn |Dα (φj − φ)| → 0 as j → ∞. Explicitly, for all α ∈ Nn and ǫ > 0 exists N such that for j ≥ N , supRn |Dα (φj − φ)| < ǫ. ∞ Lemma 0.1. For all x0 ∈ Rn and ǫ > 0 there is a function φ ∈ Cc (Rn ) such that φ(x0 ) > 0, φ ≥ 0 and supp φ ⊂ {x : |x − x0 | < ǫ}.

Proof. First one checks that the function χ defined by χ(t) = e−1/t , t > 0; χ(t) = 0, t ≤ 0, is in C ∞ (R). Then we let φ(x) = χ This has all the desired properties.
∞ An immediate corollary is that Cc (Rn ) is indeed infinite dimensional. Note also that a smooth function is not determined by its Taylor series at any point. We can now define distributions: 1

ǫ2 − |x − x0 |2 . 2

1 with x0 = 0. So | so f φ dx − f (x0 ) φ dx| ≤ |f (x) − f (x0 )| φ(x) dx ≤ |f (x0 )| 2 φ dx. for ǫ > 0 sufficiently small.2. Note that restricted to C(R ) the injectivity argument does not involve anything remotely sophisticated: if f ∈ C(Rn ).1. which is easily seen n to imply f = 0. Let δj be a sequence of positive constants with lim δj = 0. It is straightforward to check that continuity of u is equivalent to the following: ∞ for all K compact there exists m and C > 0 such that for all φ ∈ Cc (Rn ) with supp φ ⊂ K. ∞ (ii) and u is continuous.g. As an example. f φ dx = 0 as φ dx > 0. by the continuity of ψ. the map u : Cc (Rn ) → C given n 2 by u(φ) = (∂1 φ)(a) − (∂2 φ)(b). Rn The simplest distributions is the delta-distribution at some point a ∈ Rn . so one says that a sequence uj ∈ D′ (Rn ) converges to u ∈ D′ (Rn ) if u(φ) = limj→∞ uj (φ) for all ∞ φ ∈ Cc (Rn ). by the continuity of f . . The map ι : L1 (Rn ) → D′ (Rn ) is injective. n ≥ 2.2 ANDRAS VASY Definition 2. via the loc identification ι. ∞ One can also generate many similar examples. Then f defines a distribution u = ιf as follows: u(φ) = ιf (φ) = Rn f φ dx. A distribution u ∈ D′ (Rn ) is a continuous linear functional on ∞ ∞ Cc (Rn ). e. fix ǫ = 1 and let φ be as in Lemma 0. if ιf = 0 then f φ dx = 0 for all φ ∈ Cc (Rn ). A large class of distributions is obtained the following way. Then |uj (ψ) − cδ0 (ψ)| = | =| ≤ −n δj φ(x/δj )ψ(x) dx − ψ(0) −n δj φ(x/δj )ψ(0) dx| φ(x) dx| −n δj φ(x/δj )ψ(x) dx − −n δj φ(x/δj )|ψ(x) − ψ(x0 )| dx. Let uj be the distribution given by −n −n δj φ(. is locally Lebesgue integrable) – for instance. Indeed. so if φj is any sequence such that φj → φ in Cc (Rn ) then u(φj ) → u(φ) in F. Because of this lemma. Then lim uj = cδ0 . Let c = φ dx. ψ ∈ Cc (Rn ) and c ∈ F./δj ). given ǫ′ > 0. it is given by δa (φ) = φ(a). |f (x) − f (x0 )| < |f (x0 )|/2 for |x − x0 | < ǫ. u(cφ) = cu(φ) for φ. Lemma 0. Note that the integral converges since φ has compact support.e. i. (piecewise) continuous loc f would satisfy this. One usually equips D′ (Rn ) with the so-called weak-* topology. and f (x0 ) = 0 for some x0 ∈ Rn . uj (ψ) = δj φ(x/δj )ψ(x) dx. Suppose that f ∈ L1 (Rn ) (i. b ∈ R . loc ∞ Indeed.e. so |f (x) − f (x0 )| < |f (x0 )|/2 on supp φ. a distribution u is a map u : Cc (Rn ) → F such that ∞ (i) u is linear: u(φ + ψ) = u(φ) + u(ψ). there is δ ′ > 0 such that |x| < δ ′ implies |ψ(x) − ψ(x0 )| < ǫ′ . is also a distribution. |u(φ)| ≤ C |α|≤m sup |∂ α φ|. we can consider L1 (Rn ) as a subset of D′ (Rn ). That is. where a. Now let φ be as in Lemma 0.

for f ∈ C 1 (Rn ). Suppose that u ∈ D′ (R) and xu = 0. Note also that this is the only reasonable notion of a derivative as the map u → ∂j u is continuous. u(xφ) = 0 for all φ ∈ Cc (R). ∞ Suppose now that xu = 0. then H ′ (φ) = (ιH )′ (φ) = −ιH (φ′ ) = − 0 ∞ for all φ ∈ Cc (R). so the definition makes sense. and ∞ Cc (Rn ). Indeed. so (xφ)(0) = 0. if H is the Heaviside step function. δj < δ ′ . This proves our claim. This is a rather typical example. ′ ∞ δa (φ) = −δa (φ′ ) = −φ′ (a). in the integral. so certainly if x ≥ δ ′ . i. ∞ Cc (Rn ). Some examples: on R. φ ∈ Cc (R). If ψ is a test function such that ψ(0) = 0 then Taylor’s theorem allows one to write ψ = xφ with φ ∈ ∞ Cc (R) (namely φ = x−1 ψ for x = 0 extends to be C ∞ at 0). so for arbitrary u ∈ D′ (Rn ) we define gu ∈ D′ (Rn ) by gu(φ) = u(gφ).e. is dense in D′ (Rn ). Cc (Rn ) is dense in D′ (Rn ). We can now consider differentiation. Lemma 0. one can restrict the integration to |x| ≤ δj . so u(ψ) = u(xφ) = 0. where |ψ(x) − ψ(x0 )| < ǫ′ to deduce that |uj (ψ) − cδ0 (ψ)| ≤ ǫ′ −n δj φ(x/δj ) dx = cǫ′ . The idea is that we already know what the derivative of a C 1 function is. i. if x ≥ δj . and φ(x/δj ) = 0 if |x|/δj ≥ 1. in particular is continuous ∞ as a map Cc (Rn ) → C. on which we already know ∂j . Also. It is straightforward to check that ∂j u is a distribution. we proceed again by analogy with ιf where f ∈ C(Rn ).3. H(x) = 0 for x < 0. Correspondingly. so we should express it as a distribution in such a way that it obviously extends to the class of all distributions. As an example of calculation with distributions. then xu = 0 indeed: xu(φ) = u(xφ) = cδ0 (xφ) = c(xφ)(0) = 0 for all φ ∈ since x(0) = 0. In that case ∞ φ′ (x) dx = φ(0) ιf g (φ) = (f g)φ dx = f (gφ) dx = ιf (gφ). Motivated by this. Note also that distributions u may be multiplied by C ∞ functions g. Now. we make the definition: Definition 3. and it is not hard to show that one can approximate any u ∈ D′ (Rn ) in the weak-* topology by uj which are ∞ ∞ given by Cc (Rn ) functions. i. Then there is a constant c ∈ F such that u = cδ0 .e. where we used the fundamental theorem of calculus. ∞ Cc (R) . Therefore ′ H = δ0 . Proof. the distribution associated to the function ∂j f satisfies ι∂j f (φ) = ∂j f φ dx = − f ∂j φ dx = −ιf (φ) ∞ for all φ ∈ Cc (Rn ). ∞ Note that for φ ∈ gφ ∈ Cc (Rn ) since g ∈ C ∞ (Rn ). The partial derivatives of u ∈ D′ (Rn ) are defined by ∂j u(φ) = −u(∂j φ).e. so this definition makes sense. First note that if u = cδ0 .3 For j sufficiently large. uk → u implies ∂j uk → ∂j u.e. i. ∞ Note that for φ ∈ ∂j φ ∈ Cc (Rn ). so H(x) = 1 for x ≥ 0. consider the following: ∞ Cc (Rn ).

∞)t ). Note that in this case.∞)t In particular. Then by the argument of the previous paragraph. then for all u ∈ D′ (Rn ). u(x. u(ψ − αφ0 ) = 0. For instance. so f (u) can be considered as an element of D′ (R2 ). u ∈ L∞ (R2 ) (so e. ∞)t ). So we say that u ∈ L∞ (Rx × [0. One often wants solutions only for t ≥ 0. i. f (u) ∈ L∞ (R2 ) ⊂ L1 (R2 ). in which case it is helpful to reformulate the definition of a weak solution by rewriting the distributional derivatives ∞ explicitly. Rx ×[0. so L is of the form L= |α|≤m ∞ n u(φ0 ) u(φ0 ) ψ(0) = cδ0 (ψ). 0) φ(x. Let φ0 ∈ Cc (R) be such that φ0 (0) = 0. L(u) = ut + (f (u))x . In this case.4 ANDRAS VASY ∞ ∞ So now suppose that ψ ∈ Cc (R). ∞)t . for conservation laws. c = . then (ut + f (u)x )φ dx dt Rx ×[0. All these considerations go through unaffected if Rn is replaced by an open set ∞ Ω ⊂ Rn .∞)t =− Rx ×[0. (u φt + f (u)φx ) dx dt + v(x) φ(x. so loc loc the derivative makes sense.∞)t (u φt + f (u)φx ) dx dt − Rx u(x. Thus.g. u ∈ loc C(R2 ) or u has jump discontinuities) and f ∈ C(R) we say that u is a weak solution of L(u) = 0 if Lu = 0 in the sense of distributions. 0) dx = 0. let α = φ0 (0) . We ψ(0) choose α ∈ F such that ψ − αφ0 vanishes at 0. Rx Rx ×[0. it may make sense to talk about weak solutions. . ∞)t ) is a weak solution of loc ut + f (u)x = 0. If L is a linear partial differential operator. In particular. φ0 (0) φ0 (0) aα D α . Lu makes sense as an element of D′ (Rn ). then (u φt + f (u)φx ) dx dt = 0. and aα are in C (R ). u(ψ) = u((ψ − αφ0 ) + αφ0 ) = u(ψ − αφ0 ) + αu(φ0 ) = αu(φ0 ) = This finishes the proof. We say that u is a weak solution of Lu = f if Lu = f in the sense of distributions. 0) = 0 for all x. u ∈ L1 (Rn ). Suppose f ∈ L1 (Rn ). we make the following definition: Definition 4. and φ ∈ Cc (Rx × [0.e. Even if L is not linear. we consider only test functions φ ∈ Cc (Rn ) with supp φ ⊂ Ω. If u ∈ C 1 (Rx × [0. ∞)t ). 0) = v(x). if φ is supported in t > 0. L is linear with C ∞ (Rn ) loc loc coefficients.∞)t so ut + f (u)x = 0 in the sense of distributions on Rx × (0. if for all φ ∈ ∞ Cc (Rx × [0. 0) dx. so φ(x.