Partial Differential Equation

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Partial dierential equation

simple solutions for linear problems. Classic domains


where PDEs are used include acoustics, uid dynamics,
electrodynamics, and heat transfer.
A partial dierential equation (PDE) for the function
u(x1 , , xn ) is an equation of the form
)
(
u
u
2u
2u
f x1 , . . . , xn , u,
,...,
,
,...,
, . . . = 0.
x1
xn x1 x1
x1 xn
If f is a linear function of u and its derivatives, then the
PDE is called linear. Common examples of linear PDEs
include the heat equation, the wave equation, Laplaces
equation, Helmholtz equation, KleinGordon equation,
and Poissons equation.

A visualisation of a solution to the two-dimensional heat equation


with temperature represented by the third dimension

A relatively simple PDE is

In mathematics, a partial dierential equation


(PDE) is a dierential equation that contains unknown
multivariable functions and their partial derivatives. (A
special case are ordinary dierential equations (ODEs),
which deal with functions of a single variable and their
derivatives.) PDEs are used to formulate problems
involving functions of several variables, and are either
solved by hand, or used to create a relevant computer
model.

u
(x, y) = 0.
x
This relation implies that the function u(x,y) is independent of x. However, the equation gives no information on
the functions dependence on the variable y. Hence the
general solution of this equation is

PDEs can be used to describe a wide variety of phenomena such as sound, heat, electrostatics, electrodynamics,
uid dynamics, elasticity, or quantum mechanics. These u(x, y) = f (y),
seemingly distinct physical phenomena can be formalised similarly in terms of PDEs. Just as ordi- where f is an arbitrary function of y. The analogous ornary dierential equations often model one-dimensional dinary dierential equation is
dynamical systems, partial dierential equations often
model multidimensional systems. PDEs nd their generalisation in stochastic partial dierential equations.
du
(x) = 0,
dx

which has the solution

Introduction

Partial dierential equations (PDEs) are equations that


involve rates of change with respect to continuous variables. The position of a rigid body is specied by six
numbers, but the conguration of a uid is given by the
continuous distribution of several parameters, such as the
temperature, pressure, and so forth. The dynamics for the
rigid body take place in a nite-dimensional conguration
space; the dynamics for the uid occur in an innitedimensional conguration space. This distinction usually
makes PDEs much harder to solve than ordinary dierential equations (ODEs), but here again, there will be

u(x) = c,
where c is any constant value. These two examples illustrate that general solutions of ordinary dierential equations (ODEs) involve arbitrary constants, but solutions of
PDEs involve arbitrary functions. A solution of a PDE
is generally not unique; additional conditions must generally be specied on the boundary of the region where
the solution is dened. For instance, in the simple example above, the function f(y) can be determined if u is
specied on the line x = 0.
1

4 EXAMPLES

Existence and uniqueness

Although the issue of existence and uniqueness of solutions of ordinary dierential equations has a very satisfactory answer with the PicardLindelf theorem, that is
far from the case for partial dierential equations. The
CauchyKowalevski theorem states that the Cauchy problem for any partial dierential equation whose coecients are analytic in the unknown function and its derivatives, has a locally unique analytic solution. Although this
result might appear to settle the existence and uniqueness
of solutions, there are examples of linear partial dierential equations whose coecients have derivatives of all
orders (which are nevertheless not analytic) but which
have no solutions at all: see Lewy (1957). Even if the
solution of a partial dierential equation exists and is
unique, it may nevertheless have undesirable properties.
The mathematical study of these questions is usually in
the more powerful context of weak solutions.
An example of pathological behavior is the sequence (depending upon n) of Cauchy problems for the Laplace
equation

uxx =
uxy

2u
x2

2u
=
=
y x
y

u
x

)
.

Especially in physics, del or Nabla () is often used to


denote spatial derivatives, and u u
for time derivatives.
For example, the wave equation (described below) can
be written as
u
= c2 2 u
or
u
= c2 u
where is the Laplace operator.

4 Examples
4.1 Heat equation in one space dimension

2u 2u
+ 2 = 0,
x2
y

See also: Heat equation

with boundary conditions

The equation for conduction of heat in one dimension for


a homogeneous body has

u(x, 0) = 0,
ut = uxx
u
sin(nx)
(x, 0) =
,
y
n

where u(t,x) is temperature, and is a positive constant


that describes the rate of diusion. The Cauchy probwhere n is an integer. The derivative of u with respect
lem for this equation consists in specifying u(0, x)= f(x),
to y approaches 0 uniformly in x as n increases, but the
where f(x) is an arbitrary function.
solution is
General solutions of the heat equation can be found by
the method of separation of variables. Some examples
appear in the heat equation article. They are examples of
sinh(ny) sin(nx)
u(x, y) =
.
Fourier series for periodic f and Fourier transforms for
n2
non-periodic f. Using the Fourier transform, a general
This solution approaches innity if nx is not an integer solution of the heat equation has the form
multiple of for any non-zero value of y. The Cauchy
problem for the Laplace equation is called ill-posed or

not well-posed, since the solution does not continuously
2
1

u(t,
x)
=
F ()e t eix d,
depend on the data of the problem. Such ill-posed prob2
lems are not usually satisfactory for physical applications.
where F is an arbitrary function. To satisfy the initial
condition, F is given by the Fourier transform of f, that is

Notation

In PDEs, it is common to denote partial derivatives using F () = 1


2
subscripts. That is:

ux =

u
x

f (x)eix dx.

If f represents a very small but intense source of heat,


then the preceding integral can be approximated by
the delta distribution, multiplied by the strength of the

4.3

Generalised heat-like equation in one space dimension

source. For a source whose strength is normalized to 1, that are drawn backward from that point. These curves
the result is
correspond to signals that propagate with velocity m forward and backward. Conversely, the inuence of the
data at any given point on the initial line propagates with
1
the nite velocity m: there is no eect outside a triangle
F () = ,
through that point whose sides are characteristic curves.
2
This behavior is very dierent from the solution for the
and the resulting solution of the heat equation is
heat equation, where the eect of a point source appears
(with small amplitude) instantaneously at every point in

space. The solution given above is also valid if k < 0, and
2
1
the explicit formula shows that the solution depends on
u(t, x) =
e t eix d.
2
smoothly upon the data: both the forward and backward
This is a Gaussian integral. It may be evaluated to obtain Cauchy problems for the wave equation are well-posed.
(
)
1
x2
u(t, x) =
exp
.
4t
2 t

4.3 Generalised heat-like equation in one


space dimension

This result corresponds to the normal probability density Where heat-like equation means equations of the form:
for x with mean 0 and variance 2t. The heat equation
and similar diusion equations are useful tools to study
u
random phenomena.
+ f (x, t)u + g(x, t)
= Hu
t

4.2

Wave equation in one spatial dimen- where H is a SturmLiouville operator subject to the
boundary conditions:
sion

The wave equation is an equation for an unknown function u(k, x) of the form
u(x, 0) = h(x).
Then:
2

ukk = m uxx .

If:

Here u might describe the displacement of a stretched


string from equilibrium, or the dierence in air pressure
in a tube, or the magnitude of an electromagnetic eld in HXn = n Xn
a tube, and m is a number that corresponds to the velocity
Xn (a) = Xn (b) = 0
of the wave. The Cauchy problem for this equation con
sists in prescribing the initial displacement and velocity a n (t)n an (t) (Xn f (x, t), Xm )am (t) = (g(x, t), Xn )
of a string or other medium:
m
(h(x), Xn )
(Xn , Xn )

an (t)Xn (x)
u(x, t) =
an (0) =

u(0, x) = f (x),
uk (0, x) = g(x),

where f and g are arbitrary given functions. The solution where


of this problem is given by d'Alemberts formula:

1
u(k, x) = 12 [f (x mk) + f (x + mk)]+
2m

x+mk

(f, g) =
g(y) dy.

f (x)g(x)w(x) dx.
a

xmk

This formula implies that the solution at (k,x) depends on 4.4 Spherical waves
only the data on the segment of the initial line that is cut
out by the characteristic curves
Spherical waves are waves whose amplitude depends on
the radial distance r from a central point source only. For
such waves, the three-dimensional wave equation takes
x mk = constant, x + mk = constant,
the form

4 EXAMPLES
4.5.2 A typical boundary value problem

[
]
2
2
utt = c urr + ur .
r
This is equivalent to

A typical problem for Laplaces equation is to nd a solution that satises arbitrary values on the boundary of a
domain. For example, we may seek a harmonic function
that takes on the values u() on a circle of radius one. The
solution was given by Poisson:

(ru)tt = c2 [(ru)rr ] ,

2
1
1 r2
(r,
)
=
u( )d .
and hence the quantity ru satises the one-dimensional
2
2 0 1 + r 2r cos( )
wave equation. Therefore, a general solution for spherical
Petrovsky (1967, p. 248) shows how this formula can
waves has the form
be obtained by summing a Fourier series for . If r <
1, the derivatives of may be computed by dierentiating under the integral sign, and one can verify that is
1
u(t, r) = [F (r ct) + G(r + ct)] ,
analytic, even if u is continuous but not necessarily dierr
entiable. This behavior is typical for solutions of elliptic
where F and G are completely arbitrary functions. Radi- partial dierential equations: the solutions may be much
ation from an antenna corresponds to the case where G more smooth than the boundary data. This is in conis identically zero. Thus, the waveform transmitted from trast to solutions of the wave equation, and more general
an antenna has no distortion in time: the only distorting hyperbolic partial dierential equations, which typically
factor is 1/r. This feature of undistorted propagation of have no more derivatives than the data.
waves is not present if there are two spatial dimensions.

4.5

Laplace equation in two dimensions

4.6 EulerTricomi equation

The EulerTricomi equation is used in the investigation


The Laplace equation for an unknown function of two of transonic ow.
variables has the form
uxx = xuyy .
xx + yy = 0.
Solutions of Laplaces equation are called harmonic func- 4.7 Advection equation
tions.
The advection equation describes the transport of a conserved scalar in a velocity eld u = (u, v, w). It is:
4.5.1 Connection with holomorphic functions
Solutions of the Laplace equation in two dimensions are
intimately connected with analytic functions of a complex variable (a.k.a. holomorphic functions): the real and
imaginary parts of any analytic function are conjugate
harmonic functions: they both satisfy the Laplace equation, and their gradients are orthogonal. If f=u+iv, then
the CauchyRiemann equations state that

ux = vy ,

vx = uy ,

t + (u)x + (v)y + (w)z = 0.


If the velocity eld is solenoidal (that is, u = 0), then
the equation may be simplied to

t + ux + vy + wz = 0.
In the one-dimensional case where u is not constant and
is equal to , the equation is referred to as Burgers equation.

and it follows that

4.8 GinzburgLandau equation


uxx + uyy = 0,

vxx + vyy = 0.

The GinzburgLandau equation is used in modelling


superconductivity. It is

Conversely, given any harmonic function in two dimensions, it is the real part of an analytic function, at least
locally. Details are given in Laplace equation.
iut + puxx + q|u|2 u = iu

4.11

Vibrating string

where p,q C and R are constants and i is the imaginary unit.


k=

4.9

n
,
L

The Dym equation

where n is an integer. Each term in the sum corresponds


to a mode of vibration of the string. The mode with n =
The Dym equation is named for Harry Dym and occurs
1 is called the fundamental mode, and the frequencies of
in the study of solitons. It is
the other modes are all multiples of this frequency. They
form the overtone series of the string, and they are the
basis for musical acoustics. The initial conditions may
ut = u3 uxxx .
then be satised by representing f and g as innite sums
of these modes. Wind instruments typically correspond
to vibrations of an air column with one end open and one
4.10 Initial-boundary value problems
end closed. The corresponding boundary conditions are
Main article: Boundary value problem
X(0) = 0,

X (L) = 0.

Many problems of mathematical physics are formulated


as initial-boundary value problems.
The method of separation of variables can also be applied
in this case, and it leads to a series of odd overtones.

4.11 Vibrating string

The general problem of this type is solved in Sturm


Liouville theory.

If the string is stretched between two points where x=0


and x=L and u denotes the amplitude of the displacement
of the string, then u satises the one-dimensional wave 4.11.1 Vibrating membrane
equation in the region where 0 < x < L and t is unlimited. Since the string is tied down at the ends, u must also If a membrane is stretched over a curve C that forms the
satisfy the boundary conditions
boundary of a domain D in the plane, its vibrations are
governed by the wave equation
u(t, 0) = 0,

u(t, L) = 0,

as well as the initial conditions

u(0, x) = f (x),

ut (0, x) = g(x).

1
utt = uxx + uyy ,
c2
if t>0 and (x,y) is in D. The boundary condition is u(t,x,y)
= 0 if (x,y) is on C. The method of separation of variables
leads to the form

The method of separation of variables for the wave equation


u(t, x, y) = T (t)v(x, y),
utt = c2 uxx ,

which in turn must satisfy

leads to solutions of the form


1
T + k 2 T = 0,
c2
u(t, x) = T (t)X(x),
where

vxx + vyy + k 2 v = 0.

The latter equation is called the Helmholtz Equation. The


constant k must be determined to allow a non-trivial v to
satisfy the boundary condition on C. Such values of k2
T + k 2 c2 T = 0, X + k 2 X = 0,
are called the eigenvalues of the Laplacian in D, and the
where the constant k must be determined. The boundary associated solutions are the eigenfunctions of the Laplaconditions then imply that X is a multiple of sin kx, and k cian in D. The SturmLiouville theory may be extended
must have the form
to this elliptic eigenvalue problem (Jost, 2002).

CLASSIFICATION

4.12 Other examples

a second-order PDE at a given point. However, the


discriminant in a PDE is given by B 2 AC, due to the
The Schrdinger equation is a PDE at the heart of non- convention of the xy term being 2B rather than B; forrelativistic quantum mechanics. In the WKB approxima- mally, the discriminant (of the associated quadratic form)
tion it is the HamiltonJacobi equation.
is (2B)2 4AC = 4(B 2 AC), with the factor of 4
Except for the Dym equation and the GinzburgLandau dropped for simplicity.
equation, the above equations are linear in the sense that
1. B 2 AC < 0 : solutions of elliptic PDEs are as
they can be written in the form Au = f for a given linear
smooth as the coecients allow, within the interior
operator A and a given function f. Other important nonof the region where the equation and solutions are
linear equations include the NavierStokes equations dedened. For example, solutions of Laplaces equascribing the ow of uids, and Einsteins eld equations
tion are analytic within the domain where they are
of general relativity.
dened, but solutions may assume boundary valAlso, see the list of non-linear partial dierential equaues that are not smooth. The motion of a uid at
tions.
subsonic speeds can be approximated with elliptic
PDEs, and the EulerTricomi equation is elliptic
where x < 0.

Classication

Some linear, second-order partial dierential equations


can be classied as parabolic, hyperbolic and elliptic.
Others such as the EulerTricomi equation have dierent types in dierent regions. The classication provides
a guide to appropriate initial and boundary conditions,
and to the smoothness of the solutions.

5.1

Equations of rst order

Main article: First-order partial dierential equation

5.2

Linear equations of second order

Assuming uxy = uyx , the general second-order PDE in


two independent variables has the form

2. B 2 AC = 0 : equations that are parabolic at every point can be transformed into a form analogous
to the heat equation by a change of independent variables. Solutions smooth out as the transformed time
variable increases. The EulerTricomi equation has
parabolic type on the line where x = 0.
3. B 2 AC > 0 : hyperbolic equations retain any discontinuities of functions or derivatives in the initial
data. An example is the wave equation. The motion
of a uid at supersonic speeds can be approximated
with hyperbolic PDEs, and the EulerTricomi equation is hyperbolic where x > 0.
If there are n independent variables x1 , x , ..., xn, a general linear partial dierential equation of second order
has the form

Lu =
Auxx + 2Buxy + Cuyy + (lower order terms) = 0,

n
n

i=1 j=1

ai,j

2u
xi xj

terms lower-order plus = 0.

where the coecients A, B, C etc. may depend upon x The classication depends upon the signature of the
and y. If A2 + B 2 + C 2 > 0 over a region of the xy eigenvalues of the coecient matrix ai,j..
plane, the PDE is second-order in that region. This form
is analogous to the equation for a conic section:
1. Elliptic: The eigenvalues are all positive or all negative.
Ax2 + 2Bxy + Cy 2 + = 0.
More precisely, replacing x by X, and likewise for other
variables (formally this is done by a Fourier transform),
converts a constant-coecient PDE into a polynomial of
the same degree, with the top degree (a homogeneous
polynomial, here a quadratic form) being most signicant
for the classication.
Just as one classies conic sections and quadratic forms
into parabolic, hyperbolic, and elliptic based on the
discriminant B 2 4AC , the same can be done for

2. Parabolic : The eigenvalues are all positive or all


negative, save one that is zero.
3. Hyperbolic: There is only one negative eigenvalue
and all the rest are positive, or there is only one positive eigenvalue and all the rest are negative.
4. Ultrahyperbolic: There is more than one positive eigenvalue and more than one negative eigenvalue, and there are no zero eigenvalues. There is
only a limited theory for ultra-hyperbolic equations
(Courant and Hilbert, 1962).

5.4

Equations of mixed type

5.3

Systems of rst-order equations and 5.4 Equations of mixed type


characteristic surfaces

If a PDE has coecients that are not constant, it is posThe classication of partial dierential equations can be sible that it will not belong to any of these categories but
extended to systems of rst-order equations, where the rather be of mixed type. A simple but important examunknown u is now a vector with m components, and the ple is the EulerTricomi equation
coecient matrices A are m by m matrices for = 1, ...,
n. The partial dierential equation takes the form
uxx = xuyy ,
Lu =

=1

u
+ B = 0,
x

where the coecient matrices A and the vector B may


depend upon x and u. If a hypersurface S is given in the
implicit form

which is called elliptic-hyperbolic because it is elliptic


in the region x < 0, hyperbolic in the region x > 0, and
degenerate parabolic on the line x = 0.

5.5 Innite-order PDEs in quantum mechanics

(x1 , x2 , . . . , xn ) = 0,

In the phase space formulation of quantum mechanics,


where has a non-zero gradient, then S is a character- one may consider the quantum Hamiltons equations for
istic surface for the operator L at a given point if the trajectories of quantum particles. These equations are
innite-order PDEs. However, in the semiclassical excharacteristic form vanishes:
pansion, one has a nite system of ODEs at any xed order of . The evolution equation of the Wigner function
[ n
]
(
)

is also an innite-order PDE. The quantum trajectories

Q
,...,
= det
A
= 0.
are quantum characteristics, with the use of which one
x1
xn
x
=1
could calculate the evolution of the Wigner function.
The geometric interpretation of this condition is as follows: if data for u are prescribed on the surface S, then it
may be possible to determine the normal derivative of u 6 Analytical methods to solve
on S from the dierential equation. If the data on S and
PDEs
the dierential equation determine the normal derivative
of u on S, then S is non-characteristic. If the data on S
and the dierential equation do not determine the nor- 6.1 Separation of variables
mal derivative of u on S, then the surface is characteristic, and the dierential equation restricts the data on S:
Main article: Separable partial dierential equation
the dierential equation is internal to S.
1. A rst-order system Lu=0 is elliptic if no surface is
characteristic for L: the values of u on S and the
dierential equation always determine the normal
derivative of u on S.
2. A rst-order system is hyperbolic at a point if there
is a space-like surface S with normal at that point.
This means that, given any non-trivial vector orthogonal to , and a scalar multiplier , the equation
Q( + ) = 0 has m real roots 1 , 2 , ..., m. The
system is strictly hyperbolic if these roots are always distinct. The geometrical interpretation of this
condition is as follows: the characteristic form Q()
= 0 denes a cone (the normal cone) with homogeneous coordinates . In the hyperbolic case, this
cone has m sheets, and the axis = runs inside
these sheets: it does not intersect any of them. But
when displaced from the origin by , this axis intersects every sheet. In the elliptic case, the normal
cone has no real sheets.

Linear PDEs can be reduced to systems of ordinary differential equations by the important technique of separation of variables. This technique rests on a characteristic
of solutions to dierential equations: if one can nd any
solution that solves the equation and satises the boundary conditions, then it is the solution (this also applies to
ODEs). We assume as an ansatz that the dependence of
a solution on the parameters space and time can be written as a product of terms that each depend on a single
parameter, and then see if this can be made to solve the
problem.[1]
In the method of separation of variables, one reduces a
PDE to a PDE in fewer variables, which is an ordinary
dierential equation if in one variable these are in turn
easier to solve.
This is possible for simple PDEs, which are called
separable partial dierential equations, and the domain
is generally a rectangle (a product of intervals). Separable PDEs correspond to diagonal matrices thinking of

6 ANALYTICAL METHODS TO SOLVE PDES

the value for xed x" as a coordinate, each coordinate = 12 2 (T t)


can be understood separately.
v(x, ) = exp(x )u(x, ).
This generalizes to the method of characteristics, and is
also used in integral transforms.

6.5 Fundamental solution


6.2

Method of characteristics

Main article: Fundamental solution

Main article: Method of characteristics

Inhomogeneous equations can often be solved (for constant coecient PDEs, always be solved) by nding the
In special cases, one can nd characteristic curves on fundamental solution (the solution for a point source),
which the equation reduces to an ODE changing co- then taking the convolution with the boundary conditions
ordinates in the domain to straighten these curves allows to get the solution.
separation of variables, and is called the method of charThis is analogous in signal processing to understanding a
acteristics.
lter by its impulse response.
More generally, one may nd characteristic surfaces.

6.3

Integral transform

6.6 Superposition principle

Because any superposition of solutions of a linear, homoAn integral transform may transform the PDE to a sim- geneous PDE is again a solution, the particular solutions
pler one, in particular, a separable PDE. This corresponds may then be combined to obtain more general solutions.
to diagonalizing an operator.
if u1 and u2 are solutions of a homogeneous linear pde
An important example of this is Fourier analysis, which in same region R, then u= c1u1+c2u2 with any constants
diagonalizes the heat equation using the eigenbasis of si- c1 and c2 are also a solution of that pde in that same region....
nusoidal waves.
If the domain is nite or periodic, an innite sum of solutions such as a Fourier series is appropriate, but an in- 6.7 Methods for non-linear equations
tegral of solutions such as a Fourier integral is generally
required for innite domains. The solution for a point
See also the list of nonlinear partial dierential
source for the heat equation given above is an example of
equations.
the use of a Fourier integral.
There are no generally applicable methods to solve nonlinear PDEs. Still, existence and uniqueness results (such
6.4 Change of variables
as the CauchyKowalevski theorem) are often possible,
as
are proofs of important qualitative and quantitative
Often a PDE can be reduced to a simpler form with a
properties
of solutions (getting these results is a major
known solution by a suitable change of variables. For expart
of
analysis).
Computational solution to the nonlinear
ample, the BlackScholes PDE
PDEs, the split-step method, exist for specic equations
like nonlinear Schrdinger equation.
2
V
1
V
V
Nevertheless, some techniques can be used for several
+ 2 S 2 2 + rS
rV = 0
t
2
S
S
types of equations. The h-principle is the most powerful method to solve underdetermined equations. The
is reducible to the heat equation
RiquierJanet theory is an eective method for obtaining information about many analytic overdetermined systems.
u
2u
=
2

x
The method of characteristics (similarity transformation
by the change of variables (for complete details see method) can be used in some very special cases to solve
Solution of the Black Scholes Equation at the Wayback partial dierential equations.
Machine (archived April 11, 2008))

V (S, t) = Kv(x, )
(S)
x = ln K

In some cases, a PDE can be solved via perturbation analysis in which the solution is considered to be a correction
to an equation with a known solution. Alternatives are
numerical analysis techniques from simple nite dierence schemes to the more mature multigrid and nite element methods. Many interesting problems in science

7.1

Finite element method

and engineering are solved in this way using computers, 7.1 Finite element method
sometimes high performance supercomputers.
Main article: Finite element method

6.8

Lie group method

From 1870 Sophus Lie's work put the theory of dierential equations on a more satisfactory foundation. He
showed that the integration theories of the older mathematicians can, by the introduction of what are now
called Lie groups, be referred to a common source; and
that ordinary dierential equations which admit the same
innitesimal transformations present comparable diculties of integration. He also emphasized the subject of
transformations of contact.
A general approach to solving PDEs uses the symmetry property of dierential equations, the continuous
innitesimal transformations of solutions to solutions (Lie
theory). Continuous group theory, Lie algebras and
dierential geometry are used to understand the structure
of linear and nonlinear partial dierential equations for
generating integrable equations, to nd its Lax pairs, recursion operators, Bcklund transform and nally nding
exact analytic solutions to the PDE.

The nite element method (FEM) (its practical application often known as nite element analysis (FEA)) is a
numerical technique for nding approximate solutions of
partial dierential equations (PDE) as well as of integral
equations. The solution approach is based either on eliminating the dierential equation completely (steady state
problems), or rendering the PDE into an approximating
system of ordinary dierential equations, which are then
numerically integrated using standard techniques such as
Eulers method, RungeKutta, etc.

7.2 Finite dierence method


Main article: Finite dierence method
Finite-dierence methods are numerical methods for approximating the solutions to dierential equations using
nite dierence equations to approximate derivatives.

Symmetry methods have been recognized to study dif- 7.3 Finite volume method
ferential equations arising in mathematics, physics, engineering, and many other disciplines.
Main article: Finite volume method

6.9

Semianalytical methods

The adomian decomposition method, the Lyapunov articial small parameter method, and Hes homotopy perturbation method are all special cases of the more general
homotopy analysis method. These are series expansion
methods, and except for the Lyapunov method, are independent of small physical parameters as compared to the
well known perturbation theory, thus giving these methods greater exibility and solution generality.

Similar to the nite dierence method or nite element


method, values are calculated at discrete places on a
meshed geometry. Finite volume refers to the small
volume surrounding each node point on a mesh. In the
nite volume method, surface integrals in a partial differential equation that contain a divergence term are converted to volume integrals, using the divergence theorem.
These terms are then evaluated as uxes at the surfaces of
each nite volume. Because the ux entering a given volume is identical to that leaving the adjacent volume, these
methods are conservative.

Numerical methods to solve 8 See also


PDEs

Dirichlet boundary condition

The three most widely used numerical methods to solve


PDEs are the nite element method (FEM), nite volume
methods (FVM) and nite dierence methods (FDM).
The FEM has a prominent position among these methods and especially its exceptionally ecient higher-order
version hp-FEM. Other versions of FEM include the generalized nite element method (GFEM), extended nite
element method (XFEM), spectral nite element method
(SFEM), meshfree nite element method, discontinuous
Galerkin nite element method (DGFEM), Element-Free
Galerkin Method (EFGM), Interpolating Element-Free
Galerkin Method (IEFGM), etc.

Jet bundle
Laplace transform applied to dierential equations
List of dynamical systems and dierential equations
topics
Matrix dierential equation
Neumann boundary condition
Numerical partial dierential equations
Partial dierential algebraic equation

10

10

Recurrence relation
Robin boundary condition
Stochastic processes and boundary value problems

Notes

[1] Gershenfeld, Neil (2000). The nature of mathematical


modeling (Reprinted (with corr.). ed.). Cambridge: Cambridge Univ. Press. p. 27. ISBN 0521570956.

10

References

Adomian, G. (1994). Solving Frontier problems of


Physics: The decomposition method. Kluwer Academic Publishers.
Courant, R. & Hilbert, D. (1962), Methods of Mathematical Physics, II, New York: Wiley-Interscience.
Evans, L. C. (1998), Partial Dierential Equations,
Providence: American Mathematical Society, ISBN
0-8218-0772-2.
Holubov, Pavel Drbek ; Gabriela (2007). Elements of partial dierential equations ([OnlineAusg.].
ed.).
Berlin: de Gruyter.
ISBN
9783110191240.
Ibragimov, Nail H (1993), CRC Handbook of Lie
Group Analysis of Dierential Equations Vol. 1-3,
Providence: CRC-Press, ISBN 0-8493-4488-3.
John, F. (1982), Partial Dierential Equations (4th
ed.), New York: Springer-Verlag, ISBN 0-38790609-6.
Jost, J. (2002), Partial Dierential Equations, New
York: Springer-Verlag, ISBN 0-387-95428-7.
Lewy, Hans (1957), An example of a smooth linear
partial dierential equation without solution, Annals of Mathematics. Second Series, 66 (1): 155
158, doi:10.2307/1970121.
Liao, S.J. (2003), Beyond Perturbation: Introduction to the Homotopy Analysis Method, Boca Raton:
Chapman & Hall/ CRC Press, ISBN 1-58488-407X
Olver, P.J. (1995), Equivalence, Invariants and
Symmetry, Cambridge Press.
Petrovskii, I. G. (1967), Partial Dierential Equations, Philadelphia: W. B. Saunders Co..
Pinchover, Y. & Rubinstein, J. (2005), An Introduction to Partial Dierential Equations, New York:
Cambridge University Press, ISBN 0-521-84886-5.

REFERENCES

Polyanin, A. D. (2002), Handbook of Linear Partial


Dierential Equations for Engineers and Scientists,
Boca Raton: Chapman & Hall/CRC Press, ISBN 158488-299-9.
Polyanin, A. D. & Zaitsev, V. F. (2004), Handbook of Nonlinear Partial Dierential Equations,
Boca Raton: Chapman & Hall/CRC Press, ISBN 158488-355-3.
Polyanin, A. D.; Zaitsev, V. F. & Moussiaux, A.
(2002), Handbook of First Order Partial Dierential Equations, London: Taylor & Francis, ISBN 0415-27267-X.
Roubek, T. (2013), Nonlinear Partial Dierential Equations with Applications (2nd ed.), Basel,
Boston, Berlin: Birkhuser, ISBN 978-3-03480512-4, MR MR3014456
Solin, P. (2005), Partial Dierential Equations and
the Finite Element Method, Hoboken, NJ: J. Wiley &
Sons, ISBN 0-471-72070-4.
Solin, P.; Segeth, K. & Dolezel, I. (2003), HigherOrder Finite Element Methods, Boca Raton: Chapman & Hall/CRC Press, ISBN 1-58488-438-X.
Stephani, H. (1989), Dierential Equations: Their
Solution Using Symmetries. Edited by M. MacCallum, Cambridge University Press.
Wazwaz, Abdul-Majid (2009). Partial Dierential
Equations and Solitary Waves Theory. Higher Education Press. ISBN 978-3-642-00251-9.
Wazwaz, Abdul-Majid (2002). Partial Dierential Equations Methods and Applications. A.A.
Balkema. ISBN 90-5809-369-7.
Zwillinger, D. (1997), Handbook of Dierential
Equations (3rd ed.), Boston: Academic Press, ISBN
0-12-784395-7.
Gershenfeld, N. (1999), The Nature of Mathematical Modeling (1st ed.), New York: Cambridge University Press, New York, NY, USA, ISBN 0-52157095-6.
Krasilshchik, I.S. & Vinogradov, A.M., Eds.
(1999), Symmetries and Conserwation Laws for Differential Equations of Mathematical Physics, American Mathematical Society, Providence, Rhode Island,USA, ISBN 0-8218-0958-X.
Krasilshchik, I.S.; Lychagin, V.V. & Vinogradov,
A.M. (1986), Geometry of Jet Spaces and Nonlinear Partial Dierential Equations, Gordon and
Breach Science Publishers, New York, London,
Paris, Montreux, Tokyo, ISBN 2-88124-051-8.
Vinogradov, A.M. (2001), Cohomological Analysis
of Partial Dierential Equations and Secondary Calculus, American Mathematical Society, Providence,
Rhode Island,USA, ISBN 0-8218-2922-X.

11

11

Further reading

Cajori, Florian (1928). The Early History of


Partial Dierential Equations and of Partial Differentiation and Integration (PDF). The American Mathematical Monthly. 35 (9): 459467.
doi:10.2307/2298771.

12

External links

Hazewinkel, Michiel, ed. (2001), Dierential


equation, partial, Encyclopedia of Mathematics,
Springer, ISBN 978-1-55608-010-4
Partial Dierential Equations: Exact Solutions at
EqWorld: The World of Mathematical Equations.
Partial Dierential Equations: Index at EqWorld:
The World of Mathematical Equations.
Partial Dierential Equations: Methods at EqWorld:
The World of Mathematical Equations.
Example problems with solutions at exampleproblems.com
Partial
Dierential
world.wolfram.com

Equations

at

math-

Partial Dierential Equations with Mathematica


Partial Dierential Equations in Cleve Moler: Numerical Computing with MATLAB
Partial Dierential Equations at nag.com
Dispersive PDE Wiki
NEQwiki, the nonlinear equations encyclopedia
Partial dierential equation | Scholarpedia

12

13

13

TEXT AND IMAGE SOURCES, CONTRIBUTORS, AND LICENSES

Text and image sources, contributors, and licenses

13.1

Text

Partial dierential equation Source: https://en.wikipedia.org/wiki/Partial_differential_equation?oldid=744682109 Contributors: AxelBoldt, CYD, The Anome, Tarquin, Roadrunner, Nealmcb, Patrick, Michael Hardy, Oliver Pereira, Ixfd64, Ahoerstemeier, Stevenj, Silversh, Charles Matthews, Timwi, Bemoeial, Dysprosia, Jitse Niesen, Gutza, Tpbradbury, Topbanana, Robbot, MathMartin, Moink, Cbm,
Robinh, Filemon, Tosha, Giftlite, Jyril, BenFrantzDale, Michael Devore, Kmote, Waltpohl, Rpchase, PeR, Fintor, PhotoBox, L-H, Mazi,
ArnoldReinhold, Paul August, Bender235, Djordjes, Ub3rm4th, Pt, El C, David Crawshaw, Chbarts, Photonique, Crust, Msh210, Arthena,
EmmetCauleld, Jon Cates, Kbolino, Oleg Alexandrov, Tbsmith, Linas, WadeSimMiser, Mpatel, MFH, Isnow, Mandarax, Rnt20, Magister
Mathematicae, Rjwilmsi, Bertik, Andrei Polyanin, Nneonneo, R.e.b., Gseryakov, Arnero, Mathbot, Gurch, Srleer, Chobot, Eienmaru,
YurikBot, Wavelength, Borgx, Pacaro, Jschlosser, Wavesmikey, Prime Entelechy, Ojcit, FF2010, Zzuuzz, DrHok, Brian Tvedt, GraemeL,
Eigenlambda, Bjorn.sjodin, NSiDms, DStoykov, RayAYang, Nbarth, Giese, DHN-bot~enwiki, Hongooi, D.328, Yaje, Mhym, Germandemat, Donludwig, Ckatz, CyrilB, Dicklyon, Stizz, Paul Matthews, Richard77, CmdrObot, Myasuda, AndrewHowse, Cydebot, Quibik,
Dharma6662000, Thijs!bot, Egrin, Epbr123, Marek69, JustAGal, Cj67, Ben pcc, Mhaitham.shammaa, Sbarnard, Erxnmedia, The Transhumanist, Thenub314, Hut 8.5, Burga, VoABot II, JNW, David Eppstein, User A1, JaGa, Coeepusher, Marupio, R'n'B, DominiqueNC,
Manticore, Maurice Carbonaro, Foober, Salih, Belovedfreak, JonMcLoone, Policron, Cuzkatzimhut, Auent Rider, Wtt, Jmath666,
Dirkbb, Wolfrock, SieBot, Tiddly Tom, Roesser, Lagrange613, OKBot, Spartan-James, Shooke, Yhkhoo, Jonathanstray, LikeFunYouAre,
ClueBot, Rememberlands, Excirial, Winston365, Evankeane, Muro Bot, Crowsnest, Siegmaralber, Forbes72, Jss214322, PL290, Super
Cleverly, Addbot, AkhtaBot, Download, Pranagailu1436, Jasper Deng, Jarble, Luckas-bot, Yobot, OrgasGirl, Fraggle81, THEN WHO
WAS PHONE?, SwisterTwister, Wsulli74, Xqbot, Unigfjkl, Frosted14, Oscarjquintana, Amaury, FrescoBot, Aliotra, D'ohBot, Sawomir
Biay, Kwiki, Gaj0129, Pinethicket, I dream of horses, Tom.Reding, Wikiain, Iwfyita, Rausch, Orenburg1, SobakaKachalova, Pokespa,
Mandolinface, Chris in denmark, EmausBot, Dcirovic, Bkocsis, Mathsfreak, Druzhnik, Gerasime, Alpha Quadrant (alt), Donner60, ChuispastonBot, ClueBot NG, ChristophE, HappySophie, Vincius Machado Vogt, Widr, Rafnuss, Helpful Pixie Bot, J824h, Noam Duke,
Solomon7968, F=q(E+v^B), Beckman16, NotWith, Fuse809, Vanished user lt94ma34le12, YFdyh-bot, Illia Connell, BASANTDUBE,
Me, Myself, and I are Here, Hillbillyholiday, Jamesmcmahon0, Mathmensch, Leanedjuneon, JaconaFrere, Mgkrupa, Sanoonan, Pandaboy117, Treasuredwealth, KasparBot, Sonicrs, ProboscideaRubber15, Lahwaacz, Fmadd, Bear-rings and Anonymous: 239

13.2

Images

File:Commons-logo.svg Source: https://upload.wikimedia.org/wikipedia/en/4/4a/Commons-logo.svg License: CC-BY-SA-3.0 Contributors: ? Original artist: ?


File:Heat_eqn.gif Source: https://upload.wikimedia.org/wikipedia/commons/a/a9/Heat_eqn.gif License: Public domain Contributors:
This graphic was created with MATLAB. Original artist: Oleg Alexandrov
File:Navier_Stokes_Laminar.svg Source: https://upload.wikimedia.org/wikipedia/commons/7/73/Navier_Stokes_Laminar.svg License:
CC BY-SA 4.0 Contributors: Own work
Brief description of the numerical method
The following code leverages some numerical methods to simulate the solution of the 2-dimensional Navier-Stokes equation.
We choose the simplied incompressible ow Navier-Stokes Equation as follows:
<img
src='https://wikimedia.org/api/rest_v1/media/math/render/svg/1b352a66970b542690aff9810ff1514eca0952bd'
class='mwe-math-fallback-image-inline' aria-hidden='true' style='vertical-align: 2.505ex; width:27.565ex; height:6.176ex;'
alt='{\displaystyle \rho \left({\frac {\partial \mathbf {v} }{\partial t <i' />Original artist: $5+\mathbf {v} \cdot \nabla
\mathbf {v} \right)=\mu \nabla ^{2}\mathbf {v} .}">
The iterations here are based on the velocity change rate, which is given by
<img
src='https://wikimedia.org/api/rest_v1/media/math/render/svg/84351a8157ffbc3af56ed19583c97d062bfd428d'
class='mwe-math-fallback-image-inline' aria-hidden='true' style='vertical-align: 2.338ex; width:23.36ex; height:5.843ex;'
alt='{\displaystyle {\frac {\partial \mathbf {v} }{\partial t <h2' />Content license $3={\frac {\mu }{\rho }}\nabla
^{2}\mathbf {v} -\mathbf {v} \cdot \nabla \mathbf {v} .}">
Or in X coordinates:
<img
src="https://wikimedia.org/api/rest_v1/media/math/render/svg/d009b926bc255277cd55f75f9773d3721861f3ac"
class="mwe-math-fallback-image-inline aria-hidden="true style="vertical-align:
2.505ex;
width:45.497ex;
height:6.343ex;" alt="{\displaystyle {\frac {\partial v_{x}}{\partial t}}={\frac {\mu }{\rho }}({\frac {\partial
^{2}v_{x}}{\partial x^{2}}}+{\frac {\partial ^{2}v_{x}}{\partial y^{2}}})-v_{x}{\frac {\partial v_{x}}{\partial
x}}-v_{y}{\frac {\partial v_{x}}{\partial y}}.}">
The above equation gives the code. The case of Y is similar.|IkamusumeFan}}
File:Wikibooks-logo.svg Source: https://upload.wikimedia.org/wikipedia/commons/f/fa/Wikibooks-logo.svg License: CC BY-SA 3.0
Contributors: Own work Original artist: User:Bastique, User:Ramac et al.
File:Wikiquote-logo.svg Source: https://upload.wikimedia.org/wikipedia/commons/f/fa/Wikiquote-logo.svg License: Public domain
Contributors: Own work Original artist: Rei-artur
|
Creative Commons Attribution-Share Alike 3.0
}}

You might also like