See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/326668077

Solving Differential Equations by Partial Integrating Factors

Preprint · August 2017

DOI: 10.13140/RG.2.2.26747.62246

CITATIONS READS

0 877

1 author:

Tim Tarver
Bethune-Cookman University
20 PUBLICATIONS   9 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Existence and Smoothness of the Navier-Stokes Equation in Two and Three-Dimensional Euclidean Space View project

All content following this page was uploaded by Tim Tarver on 28 July 2018.

The user has requested enhancement of the downloaded file.

 Solving Differential Equations by Partial
Integrating Factors

August 30, 2017

Abstract
The field of Wave Theory is quite substantial. This field is consid-
ered something where an imagination is required to understand. We
have unknown questions in this field that need research conducted;
contributing to progression of partial differential equations and math-
ematics in general. The goal of this paper is to develop new concepts
in Wave Theory; specifically towards the Navier-Stokes equation. Part
of this concept development is showing velocity and pressure functions
to exist before moving forward in the field. Imagine a theory existing
prior to its manifestation. The missing concepts may have been found.
Lets begin by recalling a concept from ordinary differential equations.

Introduction
Recall computing integrating factors to solve ordinary differential equations.
I will yield an example of integrating factors as from any known textbook on
differential equations. We are going to find a general solution of the ordinary
differential equation
dy y
+ = x−2 .
dx x
To find the general solution to this equation, locating the integrating factor
is required. Let the integrating factor be
1
R R
P dx dx
I=e =e x = elnx = x.

Multiplying the equation by our integrating factor will give us

dy y
x + (x) = x ∗ x−2
dx x

1
 dy 1
x +y =
dx x
xy = lnx + C
lnx + C
y=
x
after integrating both sides developing our general solution.
In principle, integration factors are powerful tools for solving ordinary
differential equations. In practice of this technique, they are only found in
certain cases. Integrating factors can be found in important situations where
~u is a given function of only one variable [1]. Integrating factors are used to
solve various problems including determining exact equations [2].

Body
Let ~u(x, t) and p(x, t) be an unknown velocity vector and pressure in the set
of real numbers, R. These functions are defined for x ∈ Rn and time t ≥ 0.
Let f (x, t) be an external force applied to a fluid element creating motion
of the fluid. Let f (x, t) be equal to zero. I must show there exists smooth
functions p(x, t), ~u(x, t) ∈ R3 × [0, ∞] to satisfy conditions (1), (2), (3), (6)
and (7) of the unsolved Navier-Stokes equation. Observe a two-dimensional
partial differential equation example to be,
∂~u
+ [~u0 (x, t)]~u = xt2 .
∂x∂t
We can potentially solve one of these PDE’s using a newly defined technique.
The new concept implements same methodology for solving an ordinary dif-
ferential equation, but with partial integration. Let ~u0 (x, t) be an unknown
velocity vector defined for position x ∈ Rn and time t ≥ 0.
Let an initial velocity vector ~u0 (x, t) be equal to x2 t and treated as a
partial integrating factor in the given partial differential equation,
∂~u
+ [x2 t]~u = xt2
∂x∂t
with initial conditions ~u0 (x, 0) = ~u0 (x) = x2 (0) = 0. To create a general
solution for this partial differential equation, let our partial integrating factor
be
1 3 2
RR 2
I = e [x t]dxdt = e 6 x t .
We then multiply our given PDE with the developed integrating factor to
obtain
1 3 2 ∂~
u 1 3 2 1 3 2
e6x t + [x2 t]e 6 x t ~u = xt2 e 6 x t .
∂x∂t

2
 When we partially integrate both sides with respect to x and t, we have
the visual of
Z Z Z Z
1 3 2 ∂~u 1 3 2 1 3 2
x t 2 x t
[e 6 + [x t]e 6 ~u] = xt2 e 6 x t ∂x∂t.
∂x∂t
The left side of the equal sign becomes
Z Z
1 3 2 1 3 2
x t
e 6 ~u = xt2 e 6 x t ∂x∂t.

Notice I have not integrated the right side yet for a reason. From here,
another new concept is introduced. The new perception is called partial
integration by parts. Partial integration by parts applies to and will take
effect after the equal sign. The first step will be to partially integrate with
respect to x.
When we implement the partial integration with respect to x first, t is
the constant to look like
Z Z
1 3 2
2
t xe 6 x t ∂x∂t

Let u equal x where we take the partial derivative with respect to x giving
1 3 2 1 3 2
us ∂u
∂x
= 1. Let ∂v = e 6 x t ∂x∂t where v = x26t2 e 6 x t ∂t from the integral with
exponentials,
Z Z Z
1 3 2 1 2 2 6 1 32
e ∂x = e 6 ∂x = e[ 6 x t ]x ∂x = 2 2 e 6 x t + C
ax x t
xt
with respect to x. Notice the algebraic manipulation to the variable x where
we partially integrated the exponential, e. Let this new concept be called
partial integration with exponentials. The formula for partial integra-
tion by parts in two-dimensions will yield,
Z Z Z Z
∂v ∂u
u = uv − v .
∂x∂t ∂x∂t
We will implement partial integration by parts once with respect to t towards
obtaining a general two-dimensional equation solution. The demonstration
yields, Z Z Z Z
∂v ∂u
u = uv − v
∂x∂t ∂x∂t
Z Z Z
2 1 3 2
x t 2 6x 1 x3 t2 6 1 3 2
t xe 6 ∂x∂t = t [ 2 2 e 6 − 2 2 e 6 x t ∂t]
xt xt

3
 Z
6 1 x3 t2 6 1 3 2
= e 6 − 2 e 6 x t ∂t.
x x
From here, lets use the existing partial integration with exponentials concept
with respect to t and create,
Z
1
eat dt = eat + C
a
where a consists of a constant and one or more variables. Since we are in
partial integration, this will bring us to our solution. Then,
Z
6 1 x3 t2 6 1 3
= e 6 − 2 e[ 6 x t]t ∂t
x x
6 1 x3 t2 6 6 1 32
= e6 − 2 [ 3 e6x t ]
x x xt
6 1 32 36 1 3 2
= e6x t − 5 e6x t .
x xt
Joining together the completed partial integration of both sides obtains
1 3 2 6 1 x3 t2 36 1 3 2
e 6 x t ~u = e6 − 5 e6x t .
x xt
Now, we can solve for our general solution from the given partial differential
equation. Solving for ~u yields,
6 16 x3 t2 36 16 x3 t2
x
e − x5 t
e
~u = 1 3 2
x t
e 6

6 36
− 5
~u(x, t) =
x xt
where the function exists ∀t > 0 and ∀x 6= 0. Now, the divergence can be
shown. Finding the partial derivative of ~u with respect to x gives,
∂~u −6 180
= 2 + 6 .
∂x x xt
The divergence yields,
∞ ∞
X ∂~u X −6 180
div~u = = [ 2 + 6 ]=0
∂x x xt
(x,t)=1 (x,t)=1

where x ∈ Rn , x 6= 0 and t > 0. The divergence equaling to zero explains the

newly formed velocity vector field possessing an existing surface integral. It
also says parts of the surface have positive and negative normal components.

4
 We can now establish a pressure function using the same new concepts.
Let initial pressure p0 (x, t) be equal to 2xt as another partial integrating
factor in the given PDE,
∂p
+ [2xt]p = 2xt2 .
∂x∂t
Then, our integrating factor is
1 2 2
RR
2xtdxdt
I=e = e2x t .

Next, multiplying our integrating factor into the equation gives,

1 2 2 ∂p 1 2 2 1 2 2
e2x t
+ [2xte 2 x t ]p = 2xt2 e 2 x t .
∂x∂t
Now, we partially integrate both sides to visually obtain,
Z Z
1 2 2 1 2 2
x t
e2 p = 2xt2 e 2 x t ∂x∂t.

We have come across another partial integration by parts problem on the

right side of the equal sign. Let u = x where ∂u ∂x
= 1. Then, let ∂v =
1 2 2
x t
e 2 ∂x∂t where we partially integrate with respect to x receiving v =
2 21 x2 t2
xt2
e ∂t by the partial integration with exponentials technique,
Z Z
1 2 2 1 22
e ∂x = e[ 2 xt ]x ∂x = 2 e 2 x t + C.
ax
xt
Taking out the constant t since we are implementing partial integration
by parts suggests,
Z Z Z
2 1 2 2
x t 2 2x 1 x2 t2 2 1 x2 t2
2t xe 2 ∂x∂t = 2t [ 2 e 2 − e 2 ∂t]
xt xt2

4xt2 1 x2 t2 4t2
Z
1 2 2
= 2
e 2 − 2 e 2 x t ∂t
xt xt
Z
1 2 2
x t 4 1 2 2
= 4e 2 − e 2 x t ∂t.
x
Using the partial integration with exponentials technique with respect to t
shows, Z Z
1 2 2 1 22
e ∂t = e[ 2 x t]t ∂t = 2 e 2 x t + C.
at
xt

5
Now, finishing up with the partial integration by parts gives,
Z
1 2 2
x t 4 1 2 2
= 4e 2 − e 2 x t ∂t
x
4 2 1 22
1 2 2
= 4e 2 x t
− [ 2 e2x t ]
x xt
1 2 2 8 1 22
= 4e 2 x t − 3 e 2 x t .
xt
We bring back the left side of the equal sign, solving for our two-dimensional
general solution p,
1 2 2 1 2 2 8 1 22
e 2 x t p = 4e 2 x t − 3 e 2 x t
xt
1 2 2 8 12 x2 t2
4e 2 x t
− x3 t
e
p= 1 2 2
x t
e 2

8
p(x, t) = 4 −
x3 t
where the function exists ∀t > 0 and ∀x 6= 0.
I created a third example using another space dimension, y ∈ R3 , for
creating ~u(x, y, t) ∈ R3 as a general solution. Let the example yield an
equation such as,
∂~u
+ [2x2 y 2 t]~u = xy 2 t
∂x∂y∂t
where ~u0 (x, y, t) = 2x2 y 2 t is our initial velocity vector as a partial integrating
factor with initial conditions ~u0 (x, y, 0) = ~u0 (x, y) = 2x2 y 2 (0) = 0. Then, we
have
1 3 3 2
2x2 y 2 tdxdydt
RRR
I=e = e9x y t
1 3 3 2 ∂~u 1 3 3 2 1 3 3 2
e9x y t + [2x2 y 2 te 9 x y t ]~u = xy 2 te 9 x y t .
∂x∂y∂t
In this case, we must triple integrate both sides formulating,
Z Z Z Z Z Z
1 3 3 2
x y t ∂~u 2 2 1 3 3 2
x y t 1 3 3 2
[e 9 + [2x y te 9 ]~u] = xy 2 te 9 x y t ∂x∂y∂t
∂x∂y∂t
Z Z Z
1 3 3 2 1 3 3 2
x y t
e9 ~u = xy 2 te 9 x y t ∂x∂y∂t.

The right side of the equal sign suggests partial integration by parts to be
performed three times. The partial integration by parts case with respect to
x is Z
1 3 3 2
y t xe 9 x y t ∂x.
2

6
 1 3 3 2
Let u = x where ∂u ∂x
= 1∂y∂t. Then, let ∂v = e 9 x y t ∂x∂y∂t where the par-
1 3 3 2
tial anti-derivative with respect to x is v = x2 y93 t2 e 9 x y t ∂y∂t. The formula
yields,
Z Z Z
2 1 3 3 2
x y t 2 9 1 3 3 2
x y t 9 1 3 3 2
y t xe 9 ∂x = y t[ 3 2 e 9 − 22 e 9 x y t ∂y∂t]
xy t xt

9y 2 t 1 x3 y3 t2 9y 2 t
Z Z
1 3 3 2
= 3 2e 9 − 22 e 9 x y t ∂y∂t
xy t xt
Z Z
9 1 x3 y3 t2 9 1 3 3 2
= e9 − 2 y 2 e 9 x y t ∂y∂t.
xyt xt
Then, we compute partial integration by parts a second time by letting u = y 2
1 3 3 2 1 3 3 2
where ∂u
∂y
= 2y and ∂v = e 9 x y t ∂y∂t where v = x3 y92 t2 e 9 x y t ∂t,

9 9y 2 1 3 3 2
Z
9 1 x3 y3 t2 18y 1 3 3 2
= e9 − 2 [ 3 2 2 e9x y t − e9x y t
∂t]
xyt xt xy t x3 y 2 t2

81y 2 1 x3 y3 t2
Z
9 1 x3 y3 t2 162y 1 3 3 2
= e9 − 5 2 3 e9 + 5 2 3 e 9 x y t ∂t
xyt xy t xy t
81y 2 1 x3 y3 t2 162y
Z
9 1 x3 y3 t2 1 3 3 2
= e 9 − 5 2 3e 9 + 5 2 t−3 e 9 x y t ∂t.
xyt xy t xy
The third and final time will show our general multi-dimensional solution
with a third performance of partial integration by parts obtaining,

1 3 3 2 9 1 x3 y3 t2 81y 2 1 x3 y3 t2 162y 9 1 3 3 2 27 1 3 3 2
e9x y t
~u = e9 − 5 2 3 e9 + 5 2 [ 3 3 4 e9x y t + 3 3 5 e9x y t ]
xyt xy t xy xy t xy t

1 3 3 2
x y t 9 1 x3 y3 t2 81y 2 1 x3 y3 t2 1458 1 3 3 2 4374 1 3 3 2
e 9 ~u = e 9 − 5 2 3 e9 + 8 4 4 e9x y t + 8 4 5 e9x y t
xyt xy t xy t xy t
9 19 x3 y 3 t2 81y 2 1 3 3 2 1458 19 x3 y 3 t2 4374 19 x3 y 3 t2
xyt
e − x5 y 2 t3
e9x y t + x8 y 4 t4
e + x8 y 4 t5
e
~u = 1 3 3 2
e9x y t
9 81 1458 4374
~u(x, y, t) = − 53+ 8 44+ 8 45
xyt x t xy t xy t
where the function exists ∀t > 0 and ∀x, y 6= 0.

7
Research Implications
The new concepts developed might be able to see application towards fluid
mechanics and many other fields regarding engineering sciences. I would
like to present a possibility on bounded energy of the velocity. If we want
to establish an existing boundary of the velocity vector, a presentation of
theory must correlate. Fubini’s Theorem may have a correlation to bounded
energy on velocity vectors.

Theorem 1 (Fubini’s Theorem for Wave Theory). If u(x,y,t) is continuous

on a wave specifically in region R3 × [1, ∞), then

Z Z Z Z Z Z Z Z Z
u(x, y, t)dA = u(x, y, t)dxdydt = u(x, y, t)dydxdt
R3

where the function ~u(x, y, t) exists ∀t > 0 and ∀x, y 6= 0.

I am deciphering another implication with the velocity vector. Let the
greek letter kappa, κ, be the curvature for some known velocity vectors ~u(x, t)
and ~u(x, y, t) where ∀t > 0 and ∀x, y 6= 0 such that,

|~u0 (t) × ~u00 (t)|

κ= .
|~u0 (t)|3

8
 References
[1] Boyce, W.E. and DiPrima, R.C. (2001). Elementary Differential Equa-
tions and Boundary Value Problems. (7th ed.). New York, NY: John
Wiley and Sons, Inc.

[2] Trench, W.F. (2001). Elementary Differential Equations. Pacific Grove,

CA: Brooks/Cole Thomson Learning.

View publication stats