Preface


This section is devoted to a review of the definition of the derivative and some of its properties from the operator theory point of view.

Return to computing page for the first course APMA0330
Return to computing page for the second course APMA0340
Return to Mathematica tutorial for the first course APMA0330
Return to Mathematica tutorial for the second course APMA0340
Return to the main page for the course APMA0330
Return to the main page for the course APMA0340

Inhomogeneous Euler equations


𝔏 𝔏 𝔏 ϕ ϵ ϑ … 𝒶 𝒶 𝒶 𝛂 𝛂

Need letter: \( a \) 𝑎 𝑎 ƒ

You learn from calculus that the derivative of a smooth function f(x), defined on some interval \( (a,b) , \) is another function defined by the limit (if it exists)

\[ f'(x) \overset{\mbox{def}}{=} \frac{{\text d}f}{{\text d}x} = \lim_{h\to 0} \,\frac{f(x+h) - f(x)}{h} , \]
where we used the Lagrange and Leibniz notations for the derivative, respectively. There are known infinite many definitions of the derivative, we mention some of them:
\[ f'(x) = \lim_{h\to 0} \frac{f(x) - f(x-h)}{h} = \lim_{h\to 0} \frac{f(x+h) - f(x-h)}{2h} . \]
We will use a special notation due to L. Euler \( \texttt{D} = {\text d}/{\text d}x \quad\mbox{or} \quad\texttt{D} = {\text d}/{\text d}t \) for the derivative operator, depending on what independent variable is in use. In case of time variable t, it is a custom to utilize Newton's dot notation: \( \dot{y} = {\text d}y/{\text d}t . \) The derivative operator maps a continuously differentiable function into a continuous function. For this relation, we write \( \texttt{D} : C^1 (a,b) \to C^0 (a,b) \equiv C(a,b) \) to indicate that the derivative operator maps the set of smooth functions on some open interval into a set of continuous functions on the same interval. The derivative operator is not a bounded or continuous operator because it can transfer a bounded function into an unbounded function.

The derivative operator is a linear operator, that is,

\[ \texttt{D} \left[ u(x) + v(x) \right] = \texttt{D} \left[ u(x)\right] + \texttt{D} \left[ v(x)\right] , \qquad \texttt{D} \left[ C\,u(x)\right] = C\,\texttt{D} \left[ u(x)\right] , \]
for arbitrary constant C. We can also define integer powers of the derivative operator:
\[ \texttt{D}^0 = \texttt{I}, \qquad \texttt{D}^{n+m} = \texttt{D}^{n} \texttt{D}^{m} = \texttt{D}^{m} \texttt{D}^{n} = \frac{{\text d}^{n+m}}{{\text d} x^{n+m}} = , \]
where I is the identity operator. We would like to extend the above relation to negative indexes. So we first need to define the inverse operator, \( \texttt{D}^{-1} . \)

The inverse of the derivative is called «antiderivative» in mathematical literature mostly because it is not an operator. Why? To be an operator, \( \texttt{D}^{-1} \) should assign to every input (namely, a function) a unique output (another function). Since the kernel (or null space) of the derivative operator \( \texttt{D} \) is a one dimensional space, it assigns infinite many antiderivatives, abusing mathematicians. This follows from the identities:

\[ \texttt{D} \left[ y(x) \right] = \texttt{D} \left[ y(x) +c \right] = y' (x) \]
for any constant c. Therefore, we have infinite many inverse operators (that we mark with subscript c):
\[ \texttt{D}^{-1}_c \left[ f(x) \right] = c + \int_{x_0}^x f(s)\,{\text d}s . \]
To knock out a single inverse operator, we consider a set of continuous functions on an open interval \( (a,b) \ni x_0 \) with a prescribed condition at the point x0:
\[ C_c (a,b) = \left\{ f : f(x_0 ) = c \quad\mbox{and} \quad f \mbox{ is continuous on the interval } (a,b) \,\right\} . \] Then the derivative operator considered as a mapping
\[ \texttt{D} : C^1 (a,b) \to C_c (a,b) \subset C(a,b) \]
will be denoted by L. The latter has a unique inverse (depending on the prescribed value c): \( \texttt{D}_c^{-1} : C_c (a,b) \to C^1 (a,b) . \)

 

The operator \( \texttt{D}_c^{-1} \) is not a truly inverse, but only right inverse:
\[ \begin{split} \texttt{D}\,\texttt{D}_c^{-1} &= I \qquad \mbox{the identity operator}, \\ \texttt{D}_c^{-1} \texttt{D} &\ne I , \end{split} \]
because
\[ \texttt{D}_c^{-1} \texttt{D} f(x) = \texttt{D}_c^{-1} \, f' (x) = c + \int_{x_0}^x f' (s)\,{\text d}s = c + f(x) - f\left( x_0 \right) \ne f(x) \]
unless f(x0) = c.

Example: Let us take a constant function y(x) ≡ 1, then

\[ \texttt{D}^{-1} y (x) = \texttt{D}^{-1} \,1 = C + x , \]
where C is a constant of integration. The derivative of y(x) is identically zero, so
\[ \texttt{D}^{-1} y' (x) = \texttt{D}^{-1} \texttt{D}\,1 = \texttt{D}^{-1} 0 = C \ne 1 = \texttt{D} \, \texttt{D}^{-1} 1 = \texttt{D} \,(C+x ) . \]
So we see that \( \texttt{D} \) and \( \texttt{D}^{-1} \) do not commute (unless a condition on the constant is imposed). This means that regular indefinite integral with arbitrary constant as initial condition is only left inverse of the derivative operator.    ■

 

First order differential operator


For continuous function r(x) ∈ C and continuously differentiable function p(x) ∈ C¹, consider a first order differential operator
\[ L\left[ x, \texttt{D} \right] = r(x)\,\texttt{D}\,p(x) = r(x)\,p(x)\,\texttt{D} + r(x)\,p' (x)\, \texttt{I} = r(x) \left[ p(x)\,\texttt{D} + p' (x)\, \texttt{I} \right] , \]
where \( \texttt{I} \) is the identity operator, \( \texttt{D} \overset{\mbox{def}}{=} {\text d}/{\text d}x \) is the derivative operator, and prime denotes the derivative in Lagrange's notation. Assuming that r(x)p(x) ≠ 0, we find its inverse that depends on an arbitrary constant c:
\[ w(x) = L^{-1} \left[ x, \texttt{D} \right] f(x) = \frac{c}{p(x)} + \frac{1}{p(x)} \int_{x_0}^x \frac{f(s)}{r(s)}\,{\text d}s . \]
If the product of two functions is equal to zero at the initial point: \( r\left( x_0 \right) p\left( x_0 \right) = 0 , \) then we have a singular differential operator.

There are some examples of first order differential operators and their inverses.

\[ L\left[ x, \texttt{D} \right] = e^{-x^2} \texttt{D} e^{x^2} = \texttt{D} + 2x\,\texttt{I} , \]
with the inverse
\[ w(x) = L^{-1} \left[ x, \texttt{D} \right] f(x) = c\,e^{-x^2} + e^{-x^2} \int_0^x e^{s^2} f(s)\,{\text d}s , \]
where c is some constant. Another similar differential operator is
\[ L\left[ x, \texttt{D} \right] = e^{-x^3} \texttt{D} e^{x^3} = \texttt{D} + 3x^2\,\texttt{I} . \]
Its inverse is
\[ w(x) = L^{-1} \left[ x, \texttt{D} \right] f(x) = c\,e^{-x^3} + e^{-x^3} \int_0^x e^{s^3} f(s)\,{\text d}s . \]

 

Singular first order differential operator


If one of the functions r(x) or p(x) (or both) has a zero withing the given interval, then the linear differential operator \( L\left[ x, \texttt{D} \right] w(x) = r(x) \,\texttt{D} \left[ p(x)\,\texttt{D} w \right] \) becomes a singular one. We present some typical examples of such operators and their inverses.

Let

\[ L\left[ x, \texttt{D} \right] = \frac{1}{\tan x}\,\texttt{D}\,\tan x = \texttt{D} + \frac{1}{\sinh x\,\cosh x}\, \texttt{I} , \]
where \( \texttt{I} \) is the identity operator. Its inverse is
\[ w(x) = L^{-1} \left[ x, \texttt{D} \right] f(x) = \frac{c}{\tan x} + \frac{1}{\tan x} \int_{x_0}^x \tan s\,f(s)\,{\text d}s . \]
Another typical singular operator is
\[ L\left[ x, \texttt{D} \right] = x^{1-\gamma} \texttt{D} x^{\gamma} = x\texttt{D} + \gamma \texttt{I} \qquad \Longrightarrow \qquad w(x) = L^{-1} \left[ x, \texttt{D} \right] f(x) = c\,x^{-\gamma} + x^{-\gamma} \int_{x_0}^x s^{\gamma -1} f(s)\,{\text d}s , \]
where x0 ≠ 0.

 

Second order differential operator


Consider the general second order differential operator
\[ L\left[ x, \texttt{D} \right] y(x) = \texttt{D} \left[ p(x)\,\texttt{D} y(x) \right] + q(x)\,\texttt{D} \,y(x) + r(x)\, y = \frac{\text d}{{\text d}x} \left[ p(x)\,\frac{{\text d}y}{{\text d}x} \right] + q(x)\, \frac{{\text d}y}{{\text d}x} + r(x)\,y(x) , \]
where p(x) ∈ C¹ and other known functions q(x) and r(x) are assumed to be continuous on some fixed interval. To find the inverse operator, we multiply both sides of the equation \( L\left[ x, \texttt{D} \right] y(x) = w(x) \) by integrating factor
\[ \mu (x) = \exp \left\{ \int \frac{r(x)}{q(x)}\,{\text d}x \right\} . \]
This yields
\[ \mu (x) \,\frac{\text d}{{\text d}x} \left[ p(x)\,\frac{{\text d}y}{{\text d}x} \right] + \mu (x)\, q(x)\, \frac{{\text d}y}{{\text d}x} + \mu (x)\, r(x)\, y = \mu (x)\, w(x) . \]
Taking into accound the relation \( \mu' \, q = \mu\, r , \) we obtain
\[ \mu (x) \,\frac{\text d}{{\text d}x} \left[ p(x)\,\frac{{\text d}y}{{\text d}x} \right] + q(x) \, \frac{\text d}{{\text d}x} \left[ \mu (x)\, y \right] = \mu (x)\, w (x) , \]
so that
\[ \frac{\text d}{{\text d}x} \left[ p(x)\,\frac{{\text d}y}{{\text d}x} \right] + \frac{q(x)}{\mu (x)} \, \frac{\text d}{{\text d}x} \left[ \mu (x)\, y \right] = w(x) . \]
Suppose that the initial conditions are specified at some point x0, where the integrating factor μ(x) and its first derivative have bounded values. Upon introducing a new dependent variable z = μ y, we get
\[ \frac{\text d}{{\text d}x} \left[ p(x) \left( \frac{1}{\mu (x)} \right)' z + \frac{p(x)}{\mu (x)}\,\frac{{\text d}z}{{\text d}x} \right] + \frac{q(x)}{\mu (x)} \,\frac{{\text d}z}{{\text d}x} = w(x) . \]
If it is possible to make the coefficient of z to be a constant:
\[ p(x) \left( \frac{1}{\mu (x)} \right)' = c, \]
then our differential equation becomes
\[ \frac{\text d}{{\text d}x} \left[ s(x) \,\frac{{\text d}z}{{\text d}x} \right] + t(x)\, \frac{{\text d}z}{{\text d}x} = w(x) , \]
where
\[ s(x) = \frac{p(x)}{\mu (x)} \qquad\mbox{and}\qquad t(x) = c + \frac{q(x)}{\mu (x)} . \]
Now we multiply the above equation by another integrating factor
\[ \xi (x) = \exp \left\{ \int \frac{t(x)}{s(x)} \right\} \,{\text d}x, \]
to obtain the exact equation
\[ \frac{\text d}{{\text d}x} \left[ \xi (x)\,s(x) \,\frac{{\text d}z}{{\text d}x} \right] = \xi (x)\, w(x) . \]
Integrating with respect to x, we get
\[ \xi (x)\,s(x) \,\frac{{\text d}z}{{\text d}x} = C_1 + \int_{x_0}^x \xi (\tau )\, w(\tau )\,{\text d}\tau . \]
Second integration yields
\[ z(x) = C_2 + C_1 \int_{x_0}^x \frac{1}{\xi (x)\,s(x)}\,{\text d}x + \int_{x_0}^x \frac{1}{\xi (x)\,s(x)}\,{\text d}x \int_{x_0}^x \xi (\tau )\, w(\tau )\,{\text d}\tau . \]

 

Example: ???To be modified Consider the initial value problem

\[ x^2 \frac{{\text d}^2 y}{{\text d}x^2} + ax\,\frac{{\text d} y}{{\text d}x} + by = 3\,x^2 , \qquad y(1) , \quad y' (1) = 2. \]
We have
\[ \mu = x^{-1} , \quad \xi = x^2 , \quad h= x^3 , \quad t(x) = 2, \quad s(x) = x. \]
So z = x and y = x².    ■

 

Example: ???To be modified Consider the initial value problem

\[ \frac{{\text d}^2 y}{{\text d}x^2} - \frac{2x}{1- x^2} \, \frac{{\text d} y}{{\text d}x} + \frac{2}{1- x^2} \, y = \frac{2}{1- x^2} , \qquad y(-1) = 2 , \quad y' (-1) = -1. \]
We have
\[ \mu = x^{-1}, \quad \xi = x(1-x^2 ) , \quad h= x^2 \left( 1 - x^2 \right) , \quad t(x) = 1 - \frac{2x}{1-x^2} , \quad s(x) = x. \]
Then y = 1 - x with z = 1/x -1.    ■

 

Example: To be modified Consider the differential operator

\[ L\left[ x, \texttt{D} \right] = \texttt{D}^2 + \frac{b-x}{x}\, \texttt{D} - \frac{b}{x}\, \texttt{I} \]
Here we have p(x) = 1, q(x) = (b - x)/x, and r(x) = -b/x. Then
\[ \mu (x) = \frac{1}{b-x} , \quad \xi (x) = \frac{x^b}{b-x}\, e^{-x} , \quad h(x) = \xi (x)\, s(x) = x^b e^{-x} , \quad s(x) = b-x , \quad t(x) = \frac{(b-x)^2}{x} + 1 . \]
Then the exact solution of L y = 0 is
\[ y = \frac{1}{\mu}\,z = \frac{b-x}{b} . \]
   ■
Consider now the self-adjoint second order differential operator
\[ L\left[ x, \texttt{D} \right] y(x) = \texttt{D} \left[ p(x)\,\texttt{D} y(x) \right] - q(x)\,y(x) = \frac{\text d}{{\text d}x} \left[ p(x)\,\frac{{\text d}y}{{\text d}x} \right] - q(x)\,y(x) , \]
where p(x) and q(x) are given smooth positive functions, and \( \texttt{D}y = {\text d} y/{\text d}x \) is the derivative in the Leibniz notation. To find its inverse, we need to solve the differential equation
\[ L\left[ x, \texttt{D} \right] w(x) = f(x) . \]
For uniqueness of this operation, we have to impose two auxiliary conditions; for instance, we can take the initial conditions at some point and consider the set of continuously differentiable functions with prescribed value and its derivative:
\[ w(x_0 )= c_0 , \qquad w' (x_0 ) = c_1 . \]
Then integration gives
\[ p(x)\,\frac{{\text d}w}{{\text d}x} = p(x_0 )\,w' (x_0 ) + \int_{x_0}^x f(s)\,{\text d}s . \]
Assuming that p(x) > 0, we divide by this function and again integrate
\[ w(x) = L^{-1}\left[ x, \texttt{D} \right] f(x) = w(x_0 ) + \int_{x_0}^x \frac{{\text d}t}{p(t)} \left[ p(x_0 )\,w' (x_0 ) + \int_{x_0}^t f(s)\,{\text d}s \right] . \]
We can slightly modify the differential operator and consider
\[ L\left[ x, \texttt{D} \right] w(x) = r(x) \,\texttt{D} \left[ p(x)\,\texttt{D} w \right] = r(x)\,\frac{\text d}{{\text d}x} \left[ p(x)\,\frac{{\text d}w}{{\text d}x} \right] . \]
Assuming r(x) > 0, we make first integration
\[ p(x)\,\frac{{\text d}w}{{\text d}x} = p(x_0 )\,w' (x_0 ) + \int_{x_0}^x \frac{{\text d}t}{r(t)} \, f(t) . \]
Next integration yields the inverse operator
\[ L^{-1} \left[ x, \texttt{D} \right] f(x) = w(x_0 ) + p(x_0 )\,w' (x_0 ) \int_{x_0}^x \frac{{\text d}s}{p(s)} + \int_{x_0}^x \frac{{\text d}s}{p(s)} \int_{x_0}^s \frac{{\text d}t}{r(t)} \, f(t) \]

 

Singular second order differential operator


If one of the functions r(x) or p(x) (or both) has a zero withing the given interval, then the linear differential operator \( L\left[ x, \texttt{D} \right] w(x) = r(x) \,\texttt{D} \left[ p(x)\,\texttt{D} w \right] \) becomes a singular one. We present some typical examples of such operators and their inverses.

Let

\[ L\left[ x, \texttt{D} \right] = x\,\texttt{D} \left( \frac{1}{x}\, \texttt{D} \right) = \texttt{D}^2 - \frac{1}{x}\, \texttt{D} . \]
It is easy to verify that the above differential operator is linear:
\[ L\left[ c_1 y_1 (x) + c_2 y_2 (x) \right] = c_1 L\left[ y_1 (x) \right] + c_2 L\left[ y_2 (x) \right] , \]
for any constants c1, c2 and any two smooth functions y1(x), y2(x).

 

 

Example:

Ernst Kummer
Consider the differential equation
\[ x\, y'' + \left( \gamma -x \right) y' - \alpha\,y =0 , \]
known as confluent hypergeometric equation or Kummer's equation, introduced by Ernst Kummer in 1837. We introduce the differential operator
\[ L\left[ x, \texttt{D} \right] = x^{1-\gamma} \,\texttt{D} \left( x^{\gamma} \,\texttt{D} \right) \]
   ■

Example: because the solutions to many mathematical physics problems are expressed by the hypergeometric functions, we consider the corresponding differential equation    ■

 

We consider the nonlinear differential operator of the second order:
\[ L\left[ r, \texttt{D} \right] w(r) = \frac{1}{r} \,\texttt{D} \left[ r \left( - \texttt{D} w\right)^n \right] = \frac{1}{r} \,\frac{\text d}{{\text d}r} \left[ r \left( - \frac{{\text d}w}{{\text d}r} \right)^n \right] = -\frac{1}{r} \left( - w' \right)^{n-1} \left( w' + nr\,w'' \right) , \]
where we use the Lagrange notation for the derivative: \( w' = {\text d} w/{\text d}r . \)
Simplify[1/r*D[(r*(-D[w[r], r])^n), r]]
-(((-Derivative[1][w][r])^(-1 + n) (Derivative[1][w][r] + n r (w^\[Prime]\[Prime])[r]))/r)
Since the given differential operator is of the second order, the inverse operator will depend on two arbitrary constants; therefore, we need to consider L on a space of functions that have a specific values at a particular point and its derivative. In order to find its inverse, we need to solve the differential equation
\[ L\left[ r, \texttt{D} \right] w(r) = f(r) \qquad \Longrightarrow \qquad \frac{\text d}{{\text d}r} \left[ r \left( - \frac{{\text d}w}{{\text d}r} \right)^n \right] = r\,f(r) . \]
Integrating both sides, we get
\[ r \left( - \frac{{\text d}w}{{\text d}r} \right)^n - \left. r \left( - \frac{{\text d}w}{{\text d}r} \right)^n \right\vert_{r=r_0} = \int_{r_0}^r s\, f(s)\,{\text d}s . \]
Dividing by r and taking the n-th root, we have
\[ - \frac{{\text d}w}{{\text d}r} = \left\{ \frac{r_0}{r} \left( - w' (r_0 ) \right)^n + \frac{1}{r}\,\int_{r_0}^r s\, f(s)\,{\text d}s \right\}^{1/n} \]
Second integration yields the inverse operator
\[ L^{-1}\left[ r, \texttt{D} \right] f(r) = - \int_h^r {\text d}r \left\{ \frac{r_0}{r} \left( - w' (r_0 ) \right)^n + \frac{1}{r}\,\int_{r_0}^r s\, f(s)\,{\text d}s \right\}^{1/n} - w(h) . \]
In the above formula, you can choose r0 to be zero because the inner integral is not singular, but h should be always positive.

 

  1. Bougoffa, L., On the exact solutions for initial value problems of second-orderdifferential equations, Applied Mathematics Letters, 2009, Vol. 22, pp. 1248--1251.

 

Return to Mathematica page
Return to the main page (APMA0330)
Return to the Part 1 (Plotting)
Return to the Part 2 (First Order ODEs)
Return to the Part 3 (Numerical Methods)
Return to the Part 4 (Second and Higher Order ODEs)
Return to the Part 5 (Series and Recurrences)
Return to the Part 6 (Laplace Transform)
Return to the Part 7 (Boundary Value Problems)