Physlib

Physlib.SpaceAndTime.TimeAndSpace.ConstantTimeDist

Distributions which are constant in time

i. Overview

in this module given a distribution on `Space d`, we define the associated distribution on `Time × Space d` which is constant in time.

This is defined by integrating Schwartz Maps on `Time × Space d` over the time coordinate, to get a Schwartz Map on `Space d`.

ii. Key results

- `Space.timeIntegralSchwartz` : the integral over time of a Schwartz map on `Time × Space d` to give a Schwartz map on `Space d`. - `Space.constantTime` : the distribution on `Time × Space d` associated with a distribution on `Space d`, which is constant in time.

iii. Table of contents

- A. Properties of time integrals of Schwartz maps - A.1. Continuity as a function of space - A.2. Derivative a function of space - A.3. Differentiability as a function of space - A.4. Integrability of the derivative as a function of space - A.5. Smoothness as a function of space - B. Properties of schwartz maps at a constant space point - B.1. Integrability - B.2. Integrability of powers times norm of iterated derivatives - B.2.1. Bounds on powers times norm of iterated derivatives - B.2.2. Integrability of powers times norm of iterated derivatives - B.3. Integrability of iterated derivatives - C. Decay results for derivatives of the time integral - C.1. Moving the iterated derivative inside the time integral - C.2. Bound on the norm of iterated derivative - C.3. Bound on the norm of iterated derivative mul a power - D. The time integral as a schwartz map - E. Constant time distributions - E.1. Space derivatives of constant time distributions - E.2. Space gradient of constant time distributions - E.3. Space divergence of constant time distributions - E.4. Space curl of constant time distributions - E.5. Time derivative of constant time distributions

iv. References

A. Properties of time integrals of Schwartz maps

A.1. Continuity as a function of space

A.2. Derivative a function of space

A.3. Differentiability as a function of space

A.4. Integrability of the derivative as a function of space

A.5. Smoothness as a function of space

B. Properties of schwartz maps at a constant space point

B.1. Integrability

B.2. Integrability of powers times norm of iterated derivatives

#### B.2.1. Bounds on powers times norm of iterated derivatives

#### B.2.2. Integrability of powers times norm of iterated derivatives

B.3. Integrability of iterated derivatives

C. Decay results for derivatives of the time integral

C.1. Moving the iterated derivative inside the time integral

C.2. Bound on the norm of iterated derivative

C.3. Bound on the norm of iterated derivative mul a power

D. The time integral as a schwartz map

E. Constant time distributions

E.1. Space derivatives of constant time distributions

E.2. Space gradient of constant time distributions

E.3. Space divergence of constant time distributions

E.4. Space curl of constant time distributions

E.5. Time derivative of constant time distributions

23 declarations

theorem

The Time Integral of a Schwartz Map is Continuous

For any dimension dd and any Schwartz function η:Time×Space dR\eta : \text{Time} \times \text{Space } d \to \mathbb{R}, the function mapping each spatial point xSpace dx \in \text{Space } d to the integral of η\eta over time, xTimeη(t,x)dt, x \mapsto \int_{\text{Time}} \eta(t, x) \, dt, is continuous on Space d\text{Space } d.

theorem

Fréchet Derivative of the Time Integral of a Schwartz Map

Let dd be a natural number. Let ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) be a Schwartz function, where Time\text{Time} and Space d\text{Space } d are finite-dimensional real inner product spaces. For any point x0Space dx_0 \in \text{Space } d, the spatial function defined by the integral over time, xTimeη(t,x)dtx \mapsto \int_{\text{Time}} \eta(t, x) \, dt, is Fréchet differentiable at x0x_0. Its Fréchet derivative is given by the integral of the partial Fréchet derivative of η\eta with respect to the spatial variable: D(Timeη(t,x)dt)x=x0=TimeDxη(t,x0)dt D \left( \int_{\text{Time}} \eta(t, x) \, dt \right) \bigg|_{x=x_0} = \int_{\text{Time}} D_x \eta(t, x_0) \, dt where Dxη(t,x0)D_x \eta(t, x_0) is the Fréchet derivative of the map xη(t,x)x \mapsto \eta(t, x) evaluated at x0x_0.

theorem

The Time Integral of a Schwartz Map is Differentiable

Let dd be a natural number. For any Schwartz function ηS(Time×Space (d+1),R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } (d+1), \mathbb{R}), the function mapping each spatial point xSpace (d+1)x \in \text{Space } (d+1) to the integral of η\eta over time, xTimeη(t,x)dt, x \mapsto \int_{\text{Time}} \eta(t, x) \, dt, is differentiable on Space (d+1)\text{Space } (d+1).

theorem

The spatial derivative of a Schwartz map on Time×Space d\text{Time} \times \text{Space } d is integrable over time.

Let dd be a natural number and let η:Time×Space dR\eta: \text{Time} \times \text{Space } d \to \mathbb{R} be a Schwartz map. For any point xSpace dx \in \text{Space } d, the function mapping tTimet \in \text{Time} to the Fréchet derivative of η\eta with respect to its spatial argument at xx is integrable with respect to the volume measure. That is, the function tD2η(t,x)t \mapsto D_2 \eta(t, x) is integrable over Time\text{Time}, where D2D_2 denotes the partial Fréchet derivative with respect to the second coordinate.

theorem

The Time Integral of a Schwartz Map is CnC^n

Let dd be a natural number. For any natural number nn and any Schwartz function η:Time×Space dR\eta : \text{Time} \times \text{Space } d \to \mathbb{R}, the function defined by the integral over time, xTimeη(t,x)dt, x \mapsto \int_{\text{Time}} \eta(t, x) \, dt, is nn-times continuously differentiable (CnC^n) on Space d\text{Space } d.

theorem

Integrability of a Schwartz Map over Time at a Fixed Spatial Point

For any dimension dNd \in \mathbb{N}, let η:Time×Space(d)R\eta : \text{Time} \times \text{Space}(d) \to \mathbb{R} be a Schwartz function. For any fixed point xSpace(d)x \in \text{Space}(d), the function tη(t,x)t \mapsto \eta(t, x) is integrable with respect to the volume measure on Time\text{Time}.

theorem

(t,x)mDnη(t,x)C(1+t)rt\|(t, x)\|^m \|D^n \eta(t, x)\| \le C (1 + \|t\|)^{-r_t} for Schwartz functions η\eta

For any natural numbers n,mn, m and any spatial dimension dNd \in \mathbb{N}, there exists a natural number rtr_t such that for every Schwartz function ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) and for every point xSpace dx \in \text{Space } d, the following conditions hold: 1. The function t(1+t)rtt \mapsto (1 + \|t\|)^{-r_t} is integrable with respect to the volume measure on Time\text{Time}. 2. For all tTimet \in \text{Time}, the norm of the nn-th iterated Fréchet derivative of η\eta at (t,x)(t, x), scaled by the mm-th power of the norm of the pair (t,x)(t, x), satisfies the inequality: (t,x)mDnη(t,x)2rt+m(supjrt+m,knpj,k(η))(1+t)rt \|(t, x)\|^m \|D^n \eta(t, x)\| \leq 2^{r_t+m} \left( \sup_{j \le r_t+m, k \le n} p_{j,k}(\eta) \right) (1 + \|t\|)^{-r_t} where pj,k(η)p_{j,k}(\eta) denotes the standard Schwartz seminorm of η\eta corresponding to the jj-th power of the norm and the kk-th order derivative.

theorem

Integrability of (t,x)mDnη(t,x)\|(t, x)\|^m \|D^n \eta(t, x)\| over Time for Schwartz functions η\eta

For any dimension dNd \in \mathbb{N}, natural numbers n,mNn, m \in \mathbb{N}, and a Schwartz function ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}), the function mapping tTimet \in \text{Time} to (t,x)mDnη(t,x)\|(t, x)\|^m \|D^n \eta(t, x)\| is integrable with respect to the volume measure on Time\text{Time} for any fixed point xSpace dx \in \text{Space } d. Here, (t,x)\|(t, x)\| denotes the norm on the product space and Dnη(t,x)\|D^n \eta(t, x)\| denotes the norm of the nn-th iterated Fréchet derivative of η\eta at (t,x)(t, x).

theorem

Dnη(t,x)\|D^n \eta(t, x)\| is integrable over Time for Schwartz functions η\eta

For any spatial dimension dNd \in \mathbb{N} and any natural number nNn \in \mathbb{N}, let ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) be a Schwartz function. For any fixed point xSpace dx \in \text{Space } d, the function mapping tTimet \in \text{Time} to the norm of the nn-th iterated Fréchet derivative of η\eta at (t,x)(t, x), denoted by Dnη(t,x)\|D^n \eta(t, x)\|, is integrable with respect to the volume measure on Time\text{Time}.

theorem

Dnη(t,x)D^n \eta(t, x) is integrable over Time\text{Time} for Schwartz functions η\eta

For any spatial dimension dNd \in \mathbb{N}, any natural number nNn \in \mathbb{N}, and any Schwartz function ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}), the function mapping tTimet \in \text{Time} to the nn-th iterated Fréchet derivative of η\eta at (t,x)(t, x), denoted by Dnη(t,x)D^n \eta(t, x), is integrable with respect to the volume measure on Time\text{Time} for any fixed point xSpace dx \in \text{Space } d.

theorem

The nn-th Fréchet Derivative of a Time Integral Equals the Integral of the nn-th Spatial Derivative

Let dd be a natural number and let η:Time×Space dR\eta: \text{Time} \times \text{Space } d \to \mathbb{R} be a Schwartz map. For any nNn \in \mathbb{N}, the nn-th Fréchet derivative of the spatial function defined by the integral over time, xTimeη(t,x)dtx \mapsto \int_{\text{Time}} \eta(t, x) \, dt, evaluated at xSpace dx \in \text{Space } d and applied to a sequence of vectors y1,,ynSpace dy_1, \dots, y_n \in \text{Space } d, is equal to the integral of the nn-th Fréchet derivative of η\eta evaluated at (t,x)(t, x) and applied to the vectors ((0,y1),,(0,yn))((0, y_1), \dots, (0, y_n)). That is: Dn(Timeη(t,x)dt)(y1,,yn)=TimeDnη(t,x)((0,y1),,(0,yn))dt D^n \left( \int_{\text{Time}} \eta(t, x) \, dt \right) (y_1, \dots, y_n) = \int_{\text{Time}} D^n \eta(t, x) ((0, y_1), \dots, (0, y_n)) \, dt where DnD^n denotes the nn-th Fréchet derivative.

theorem

The nn-th Fréchet Derivative of a Time Integral Equals the Integral of nn-th Spatial Derivatives

Let dNd \in \mathbb{N} be the spatial dimension and nNn \in \mathbb{N} the order of differentiation. For any Schwartz function ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) and any point xSpace dx \in \text{Space } d, the nn-th Fréchet derivative of the spatial function xTimeη(t,x)dtx \mapsto \int_{\text{Time}} \eta(t, x) \, dt is equal to the integral over Time\text{Time} of the nn-th Fréchet derivative of η\eta at (t,x)(t, x) pre-composed with the linear inclusion L:Space dTime×Space dL: \text{Space } d \to \text{Time} \times \text{Space } d given by y(0,y)y \mapsto (0, y) in each of its nn arguments. Specifically: Dn(Timeη(t,x)dt)=Time(Dnη(t,x)(L,,L))dt D^n \left( \int_{\text{Time}} \eta(t, x) \, dt \right) = \int_{\text{Time}} \left( D^n \eta(t, x) \circ (L, \dots, L) \right) \, dt where DnD^n denotes the nn-th Fréchet derivative.

theorem

Dnηdt(Dnηdt)Ln\| D^n \int \eta \, dt \| \le (\int \| D^n \eta \| \, dt) \cdot \| L \|^n

Let dNd \in \mathbb{N} be the spatial dimension and nNn \in \mathbb{N} the order of differentiation. For any Schwartz function ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) and any point xSpace dx \in \text{Space } d, the norm of the nn-th Fréchet derivative of the spatial function defined by the integral over time, xTimeη(t,x)dtx \mapsto \int_{\text{Time}} \eta(t, x) \, dt, is bounded by: Dn(Timeη(t,x)dt)(TimeDnη(t,x)dt)Ln \left\| D^n \left( \int_{\text{Time}} \eta(t, x) \, dt \right) \right\| \leq \left( \int_{\text{Time}} \| D^n \eta(t, x) \| \, dt \right) \cdot \| L \|^n where DnD^n denotes the nn-th Fréchet derivative and L:Space dTime×Space dL: \text{Space } d \to \text{Time} \times \text{Space } d is the continuous linear inclusion map defined by L(y)=(0,y)L(y) = (0, y).

theorem

xmDnηdt\|x\|^m \|D^n \int \eta \, dt\| is bounded by a supremum of Schwartz seminorms of η\eta

For any spatial dimension dNd \in \mathbb{N} and any natural numbers n,mNn, m \in \mathbb{N}, there exists a natural number rtr_t such that for every Schwartz function ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) and for every spatial point xSpace dx \in \text{Space } d, the following conditions hold: 1. The function t(1+t)rtt \mapsto (1 + \|t\|)^{-r_t} is integrable with respect to the volume measure on Time\text{Time}. 2. The product of xm\|x\|^m and the norm of the nn-th Fréchet derivative of the time integral xTimeη(t,x)dtx \mapsto \int_{\text{Time}} \eta(t, x) \, dt satisfies the inequality: xmDn(Timeη(t,x)dt)(Time1(1+t)rtdt)Ln2rt+m(supjrt+m,knpj,k(η)) \|x\|^m \left\| D^n \left( \int_{\text{Time}} \eta(t, x) \, dt \right) \right\| \leq \left( \int_{\text{Time}} \frac{1}{(1 + \|t\|)^{r_t}} \, dt \right) \|L\|^n \cdot 2^{r_t+m} \left( \sup_{j \leq r_t+m, k \leq n} p_{j,k}(\eta) \right) where DnD^n denotes the nn-th Fréchet derivative, L:Space dTime×Space dL: \text{Space } d \to \text{Time} \times \text{Space } d is the continuous linear map defined by L(y)=(0,y)L(y) = (0, y), and pj,k(η)p_{j,k}(\eta) denotes the standard Schwartz seminorm of η\eta associated with the jj-th power of the norm and the kk-th order derivative.

definition

Time integration of Schwartz maps

For any dimension dd, the continuous linear map timeIntegralSchwartz\text{timeIntegralSchwartz} maps a Schwartz function ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) to a Schwartz function in S(Space d,R)\mathcal{S}(\text{Space } d, \mathbb{R}) by integrating over the time coordinate. For an element xSpace dx \in \text{Space } d, the resulting function is defined as: (timeIntegralSchwartzη)(x)=Timeη(t,x)dt (\text{timeIntegralSchwartz} \, \eta)(x) = \int_{\text{Time}} \eta(t, x) \, dt where the integration is performed with respect to the volume measure on the Time\text{Time} space.

theorem

Pointwise evaluation of the time integral of a Schwartz function

For any dimension dd, let ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) be a Schwartz function and xSpace dx \in \text{Space } d be a point in space. The value of the time-integrated Schwartz function timeIntegralSchwartzη\text{timeIntegralSchwartz} \, \eta at the point xx is given by the integral of η\eta over the time coordinate: (timeIntegralSchwartzη)(x)=Timeη(t,x)dt (\text{timeIntegralSchwartz} \, \eta)(x) = \int_{\text{Time}} \eta(t, x) \, dt where the integration is performed with respect to the volume measure on Time\text{Time}.

definition

Time-constant distribution associated with a spatial distribution

The linear map `constantTime` transforms an MM-valued distribution ff on Space d\text{Space } d into a distribution on Time×Space d\text{Time} \times \text{Space } d that is constant in time. For any Schwartz test function ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}), the action of the resulting distribution is defined by: (constantTimef)(η)=f(xTimeη(t,x)dt) (\text{constantTime} \, f)(\eta) = f \left( x \mapsto \int_{\text{Time}} \eta(t, x) \, dt \right) In other words, the distribution ff is applied to the time-integral of the test function η\eta.

theorem

(constantTime f)(η)=f(timeIntegralSchwartz η)(\text{constantTime } f)(\eta) = f(\text{timeIntegralSchwartz } \eta)

For any dimension dd and any MM-valued distribution ff on Space d\text{Space } d (where MM is a normed space over R\mathbb{R}), let ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) be a Schwartz test function. The action of the time-constant distribution constantTime f\text{constantTime } f on η\eta is given by applying ff to the time-integral of η\eta: (constantTime f)(η)=f(timeIntegralSchwartz η) (\text{constantTime } f)(\eta) = f(\text{timeIntegralSchwartz } \eta) where timeIntegralSchwartz η\text{timeIntegralSchwartz } \eta is the Schwartz function on Space d\text{Space } d defined by xTimeη(t,x)dtx \mapsto \int_{\text{Time}} \eta(t, x) \, dt.

theorem

xi(constantTime f)=constantTime (fxi)\frac{\partial}{\partial x_i} (\text{constantTime } f) = \text{constantTime } (\frac{\partial f}{\partial x_i})

Let MM be a real normed vector space and dd be a natural number. For any MM-valued distribution ff on Space d\text{Space } d and any spatial coordinate index i{0,,d1}i \in \{0, \dots, d-1\}, the ii-th spatial partial derivative of the time-constant distribution associated with ff is equal to the time-constant distribution associated with the ii-th partial derivative of ff. Specifically, xi(constantTime f)=constantTime(fxi) \frac{\partial}{\partial x_i} (\text{constantTime } f) = \text{constantTime} \left( \frac{\partial f}{\partial x_i} \right) where constantTime f\text{constantTime } f is the distribution on Time×Space d\text{Time} \times \text{Space } d whose action on a Schwartz test function η(t,x)\eta(t, x) is given by f(xTimeη(t,x)dt)f(x \mapsto \int_{\text{Time}} \eta(t, x) \, dt).

theorem

space(constantTime f)=constantTime(f)\nabla_{\text{space}} (\text{constantTime } f) = \text{constantTime} (\nabla f)

For any natural number dd, let ff be a real-valued distribution on Space d\text{Space } d. The spatial gradient of the time-constant distribution associated with ff is equal to the time-constant distribution associated with the distributional gradient of ff. That is, space(constantTime f)=constantTime(f) \nabla_{\text{space}} (\text{constantTime } f) = \text{constantTime} (\nabla f) where constantTime f\text{constantTime } f is the distribution on Time×Space d\text{Time} \times \text{Space } d whose action on a Schwartz test function η(t,x)\eta(t, x) is f(xTimeη(t,x)dt)f(x \mapsto \int_{\text{Time}} \eta(t, x) \, dt), space\nabla_{\text{space}} is the gradient with respect to the spatial coordinates, and f\nabla f is the distributional gradient of ff on Space d\text{Space } d.

theorem

space(constantTime f)=constantTime(f)\nabla_{\text{space}} \cdot (\text{constantTime } \mathbf{f}) = \text{constantTime} (\nabla \cdot \mathbf{f})

Let dd be a natural number representing the spatial dimension. For any vector-valued distribution f\mathbf{f} on Space d\text{Space } d (where Space dRd\text{Space } d \cong \mathbb{R}^d) taking values in the Euclidean space Rd\mathbb{R}^d, let constantTime f\text{constantTime } \mathbf{f} be the distribution on Time×Space d\text{Time} \times \text{Space } d associated with f\mathbf{f} that is constant in time. The spatial divergence of constantTime f\text{constantTime } \mathbf{f} is equal to the time-constant distribution associated with the distributional divergence of f\mathbf{f}: space(constantTime f)=constantTime(f) \nabla_{\text{space}} \cdot (\text{constantTime } \mathbf{f}) = \text{constantTime} (\nabla \cdot \mathbf{f}) where space\nabla_{\text{space}} \cdot denotes the divergence operator acting on the spatial coordinates of a distribution on Time×Space d\text{Time} \times \text{Space } d, and f\nabla \cdot \mathbf{f} is the divergence of the distribution f\mathbf{f} on Space d\text{Space } d.

theorem

space×(constantTime f)=constantTime (×f)\nabla_{\text{space}} \times (\text{constantTime } f) = \text{constantTime } (\nabla \times f)

Let ff be an R3\mathbb{R}^3-valued distribution on Space 3\text{Space } 3. Let constantTime f\text{constantTime } f be the distribution on Time×Space 3\text{Time} \times \text{Space } 3 defined by its action on a Schwartz test function η(t,x)\eta(t, x) as (constantTime f)(η)=f(xTimeη(t,x)dt)(\text{constantTime } f)(\eta) = f(x \mapsto \int_{\text{Time}} \eta(t, x) \, dt). Then the spatial distributional curl of constantTime f\text{constantTime } f is equal to the time-constant distribution associated with the curl of ff: space×(constantTime f)=constantTime (×f) \nabla_{\text{space}} \times (\text{constantTime } f) = \text{constantTime } (\nabla \times f) where space×\nabla_{\text{space}} \times denotes the curl operator acting on the spatial coordinates of distributions on Time×Space 3\text{Time} \times \text{Space } 3, and ×f\nabla \times f is the curl of the distribution ff on Space 3\text{Space } 3.

theorem

t(constantTime f)=0\frac{\partial}{\partial t} (\text{constantTime } f) = 0

Let MM be a real normed space and dNd \in \mathbb{N} be the spatial dimension. For any MM-valued distribution ff on Space d\text{Space } d, let constantTime f\text{constantTime } f denote the distribution on Time×Space d\text{Time} \times \text{Space } d associated with ff which is constant in time. This distribution is defined by its action on a Schwartz test function ηS(Time×Space d,R)\eta \in \mathcal{S}(\text{Time} \times \text{Space } d, \mathbb{R}) as (constantTime f)(η)=f(xTimeη(t,x)dt)(\text{constantTime } f)(\eta) = f(x \mapsto \int_{\text{Time}} \eta(t, x) \, dt). Then, the temporal derivative of this distribution is the zero distribution: t(constantTime f)=0 \frac{\partial}{\partial t} (\text{constantTime } f) = 0