Problems
  1. For any integer k0k \geq 0, define a linear functional on C0(R)C_0^\infty (\mathbb{R}):
    δ(k),φ=(1)kφ(k)(0),\langle \delta^{(k)}, \varphi \rangle = (-1)^k \varphi^{(k)}(0),
    where φ(k)\varphi^{(k)} denotes the kk-th derivative. Prove that this defines a distribution on R\mathbb{R} and that its order is precisely kk.

  2. Define a linear functional on C0(R)C_0^\infty (\mathbb{R}):
    u,φ=k=0φ(k)(0).\langle u, \varphi \rangle = \sum_{k=0}^{\infty} \varphi^{(k)}(0).
    Prove that this defines a distribution on R\mathbb{R} but we cannot define its order.

  3. Prove that the order of vp1x\text{vp} \, \frac{1}{x} is 1.

  4. Prove that if the test function φC0(R)\varphi \in C_0^\infty (\mathbb{R}) satisfies φ(0)=0\varphi(0) = 0, then
    vp1x,φ=Rφ(x)φ(0)xdx.\left\langle \text{vp} \, \frac{1}{x}, \varphi \right\rangle = \int_{\mathbb{R}} \frac{\varphi(x) - \varphi(0)}{x} \, dx.

  5. Prove that the function on R>0\mathbb{R}_{>0}
    f(x)=e1x1(0,+)(x)f(x) = e^{\frac{1}{x}} \mathbf{1}_{(0,+\infty)}(x)
    cannot be extended to a distribution on R\mathbb{R}, i.e., there does not exist uD(R)u \in \mathcal{D}'(\mathbb{R}) such that for any φD((0,))\varphi \in \mathcal{D}((0, \infty)),
    u,φ=01xφ(x)dx.\langle u, \varphi \rangle = \int_0^{\infty} \frac{1}{x} \varphi(x) \, dx.

  6. For any ε>0\varepsilon > 0, define
    fε(x)=1x1(,ε)(ε,+)(x).f_\varepsilon (x) = \frac{1}{x} \mathbf{1}_{(-\infty, -\varepsilon) \cup (\varepsilon, +\infty)}(x).
    Prove that as ε0\varepsilon \to 0, fεf_\varepsilon has a limit in the sense of distributions and compute this limit.

  7. For any ε>0\varepsilon > 0, define
    gε(x)=1x1(,ε2)(ε2,+)(x).g_\varepsilon (x) = \frac{1}{x} \mathbf{1}_{(-\infty, -\varepsilon^2) \cup (\varepsilon^2, +\infty)}(x).
    Consider these as a family of distributions in D(R)\mathcal{D}'(\mathbb{R}). Prove that as ε0\varepsilon \to 0, gεg_\varepsilon does not have a limit in the sense of distributions.

  8. Suppose μ\mu is a probability measure on (R,B(R))(\mathbb{R}, \mathcal{B}(\mathbb{R})), i.e., μ(R)=1\mu(\mathbb{R}) = 1. Let f(x)=μ((,x))f(x) = \mu((-\infty, x)), prove that in the sense of distributions
    f=D(R)=μ.f' = \mathcal{D}'(\mathbb{R}) = \mu.

  9. Conversely, given a monotonically increasing function f(x)f(x) on R\mathbb{R} such that
    limxf(x)=0,limx+f(x)=1,\lim_{x \to -\infty} f(x) = 0, \quad \lim_{x \to +\infty} f(x) = 1,
    prove that its derivative in the sense of distributions ff' can be viewed as a probability measure on (R,B(R))(\mathbb{R}, \mathcal{B}(\mathbb{R})), i.e., there exists a probability measure μ\mu such that for any φD(R)\varphi \in \mathcal{D}(\mathbb{R}),
    f,φ=Rφdμ.\langle f', \varphi \rangle = \int_{\mathbb{R}} \varphi \, d\mu.

  10. For αR\alpha \in \mathbb{R}, define the function on R\mathbb{R}
    x+α={xα,x>0;0,x0.x_+^{\alpha} = \begin{cases} x^{\alpha}, & x > 0; \\ 0, & x \leq 0. \end{cases}
    Clearly, when α>1\alpha > -1, x+αx_+^{\alpha} is locally integrable, thus giving a distribution on R\mathbb{R}. When α>0\alpha > 0, prove
    (x+α)=α(x+α1).(x_+^{\alpha})' = \alpha (x_+^{\alpha-1}).

  11. If α\alpha is a positive integer, as a distribution on R\mathbb{R} (and as the derivative of a distribution), prove
    (x+α)(α)=α!H(x).(x_+^{\alpha})^{(\alpha)} = \alpha! H(x).
    Here H(x)H(x) is the Heaviside function.

  12. Now assume 2<α<1-2 < \alpha < -1, for any φC0(R)\varphi \in C_0^\infty (\mathbb{R}) and any ε>0\varepsilon > 0, prove
    εxαφ(x)dx=(α+1)1φ(ε)εα+1(α+1)1εxα+1φ(x)dx.\int_{\varepsilon}^{\infty} x^{\alpha} \varphi(x) \, dx = -(\alpha + 1)^{-1} \varphi(\varepsilon) \varepsilon^{\alpha + 1} - (\alpha + 1)^{-1} \int_{\varepsilon}^{\infty} x^{\alpha + 1} \varphi'(x) \, dx.
    This defines the distribution
    x+α,φ=limε0(α+1)1εxα+1φ(x)dx=(α+1)1(x+α+1),φ.\langle x_+^{\alpha}, \varphi \rangle = \lim_{\varepsilon \to 0} (\alpha + 1)^{-1} \int_{\varepsilon}^{\infty} x^{\alpha + 1} \varphi'(x) \, dx = (\alpha + 1)^{-1} \langle (x_+^{\alpha + 1})', \varphi \rangle.
    For general non-integer α<1\alpha < -1, a similar definition can be made, satisfying the relation in Question 10.

  13. Find all uD(R)u \in \mathcal{D}'(\mathbb{R}) such that in the sense of distributions
    xu=0.x \cdot u = 0.

  14. Find all distributions uu on R\mathbb{R} such that
    xu=δ0.x \cdot u = \delta_0.

  15. Prove that for each uD(R)u \in \mathcal{D}'(\mathbb{R}), there exists u~D(R)\tilde{u} \in \mathcal{D}'(\mathbb{R}) such that in the sense of distributions
    xu~=u.x \cdot \tilde{u} = u.

  16. Find all distributions uu on R\mathbb{R} such that
    xu=1.x \cdot u = 1.

  17. Find all distributions uu on R\mathbb{R} such that
    xu=pv1x.x \cdot u = \text{pv} \, \frac{1}{x}.

  18. Find all distributions uu on R\mathbb{R} such that
    xu=x+α,x \cdot u = x_+^{\alpha},
    where α<1\alpha < -1 and is not an integer.

  19. Find all uD(R)u \in \mathcal{D}'(\mathbb{R}) such that in the sense of distributions
    u+xu=0.u + x \cdot u' = 0.

  20. For a test function φC0(Rn)\varphi \in C_0^\infty(\mathbb{R}^n) and a real number λ0\lambda \neq 0, define
    φλ(x)=φ(xλ).\varphi_{\lambda}(x) = \varphi \left( \frac{x}{\lambda} \right).
    For a distribution uD(Rn)u \in \mathcal{D}'(\mathbb{R}^n), define the distribution uλu_{\lambda} as
    uλ,φ=u,λnφλ.\langle u_{\lambda}, \varphi \rangle = \langle u, |\lambda|^n \varphi_{\lambda} \rangle.
    If for each λ>0\lambda > 0,
    uλ=Dλdu,u_{\lambda} \overset{\mathcal{D}'}{=} \lambda^d u,
    then uu is called a homogeneous distribution of degree dd, where dRd \in \mathbb{R}. Prove that vp1x\text{vp} \, \frac{1}{x} and x+αx_+^{\alpha} (where α<1\alpha < -1 and is not an integer) are homogeneous distributions on D(R)\mathcal{D}'(\mathbb{R}) and determine their degree.

  21. Prove that δ0\delta_0 is a homogeneous distribution on D(Rn)\mathcal{D}'(\mathbb{R}^n) and determine its degree.

  22. Assume uD(Rn)u \in \mathcal{D}'(\mathbb{R}^n) is a homogeneous distribution of degree dd, prove that xiuD(Rn)\partial_{x_i} u \in \mathcal{D}'(\mathbb{R}^n) is a homogeneous distribution of degree d1d - 1.

  23. Given NN homogeneous distributions u1,u2,,uND(Rn)u_1, u_2, \dots, u_N \in \mathcal{D}'(\mathbb{R}^n) with distinct degrees, prove they are linearly independent in D(Rn)\mathcal{D}'(\mathbb{R}^n).

  24. Given a test function φ0D(Rn)\varphi_0 \in \mathcal{D}(\mathbb{R}^n), a positive real number λ\lambda and λ0\lambda_0, define
    φλ(x)=φλ(x)φλ0(x)λλ0.\varphi_{\lambda}(x) = \frac{\varphi_{\lambda}(x) - \varphi_{\lambda_0}(x)}{\lambda - \lambda_0}.
    Prove that as λλ0\lambda \to \lambda_0, the function above has a limit in D(Rn)\mathcal{D}(\mathbb{R}^n) and calculate this limit.

  25. (Euler’s Formula for Homogeneous Distributions) Given uD(Rn)u \in \mathcal{D}'(\mathbb{R}^n), dRd \in \mathbb{R}. Prove that uu is a homogeneous distribution of degree dd if and only if
    k=1nxkuxk=du.\sum_{k=1}^n x_k \frac{\partial u}{\partial x_k} = d \cdot u.

  26. Find all homogeneous distributions of degree 0 and 1 in D(R1)\mathcal{D}'(\mathbb{R}^1).