You are on page 1of 20

Polytechnic University of the Philippines

College of Engineering

Lecturer: Federico A. Roy Jr, BSEE, MEE, PhD, IIEE (member)


MATH 20063
Differential
Equations
Lecture 6

5.1 Integrating Factors by Inspection

The Objective of this lesson is to enable to find integrating factors by


inspection. To be able to do this, the following common exact differentials
must be easily recognized.
(1) 𝑑 𝑥𝑦 = 𝑥𝑑𝑦 + 𝑦𝑑𝑥
𝑥
(2) 𝑑 𝑦𝑑𝑥$𝑥𝑑𝑦
𝑦 =
𝑦2
(3) 𝑑 𝑦
𝑥𝑑𝑦$𝑦𝑑𝑥
𝑥 =
𝑥2

(4) 𝑑 arctan 𝑥𝑑𝑦$𝑦𝑑𝑥


𝑦 =
𝑥2%𝑦2
𝑥
Example 5.◤1 Find a set of solution for the d.e. 𝑦 𝑦 2 + 1 𝑑𝑥 + x 𝑦 2 − 1 𝑑𝑦 = 0

Solutions: Given the d.e 𝑦 𝑦 2 + 1 𝑑𝑥 + x 𝑦 2 − 1 𝑑𝑦 = 0 − −(1)

Rearranging and simplifying (1) will result to

𝑦 2 𝑦𝑑𝑥 + 𝑥𝑑𝑦 + 𝑦𝑑𝑥 − 𝑥𝑑𝑦 = 0 − − 2 verify this!

Further simplification of (2) will be 𝑦𝑑𝑥 + 𝑥𝑑𝑦 + 𝑦𝑑𝑥%𝑥𝑑𝑦


𝑦
2 = 0 − −(3)

𝑥
Observing (3), we can deduce that 𝑑 𝑥𝑦 = 𝑦𝑑𝑥 + 𝑥𝑑𝑦 and 𝑑
𝑦 𝑦𝑑𝑥%𝑥𝑑𝑦
=
𝑦2

by product rule and quotient rule respectively , hence (3) , can now be written as

𝑥
𝑑 𝑥𝑦 + 𝑑 𝑥
𝑦 = 0 and integrating it will ∫ 𝑑 𝑥𝑦 + ∫ 𝑑 = ∫ 0 will give
𝑦

𝑥 𝑦 2 + 1 = 𝑐𝑦
𝑥𝑦 + 𝑥
𝑦 = 𝑐 or ⟶ general solution of (1)

Example 5.1a. Solve the d.e. in Example 5.1 using a different method

Solutions: Given the d.e. 𝑦 𝑦 2 + 1 𝑑𝑥 + x 𝑦 2 − 1 𝑑𝑦 = 0 − −(1)

𝑥𝑑𝑦
Rearranging and simplifying (1) will result to &%𝑦2 𝑑𝑥 &%𝑦2
𝑑𝑥 = and = 𝑑𝑦—(2)
𝑦 𝑦 2 '& 𝑥 𝑦 �
�2'&

𝑑𝑥
using separation of variables, integrating (2) (&%𝑦2)
∫ 𝑥 =∫ 𝑑𝑦 − − 3
𝑦 𝑦 2 '&

Using partial integration of the right hand side of (3), the results will be

𝑑𝑥
∫ & 2𝑦 ) − −(4) verify this!
𝑥 = ∫(𝑦 − 𝑦 2 '&

and performing integration in (4), which when integrated will result to

ln 𝑥 = ln 𝑦 − ln 𝑦 2 + 1 + 𝑙𝑛𝑐 , finally equal to


which is consistent with the 𝑥 𝑦 2 + 1 = 𝑐𝑦
answer in Example 5.1. Which of the two methods is preferable? Why?

Exercise 5.1

(a) Find the general solution of the given d.e.(s)


𝑑𝑦
(1)
𝑑𝑥 − 𝑚𝑦 = 𝐶$𝑒𝑚𝑥 , where 𝐶$ and 𝑚 are constants Ans. 𝑦 = 𝑐$𝑥 + 𝑐& 𝑒𝑚𝑥
(2) 𝑑𝑥 − 1 + 2𝑥𝑡𝑎𝑛𝑦 𝑑𝑦 = 0 Ans. 2𝑥 𝑐𝑜𝑠&𝑦 = 𝑦 + 𝑐 + 𝑠𝑖𝑛𝑦 𝑐𝑜𝑠𝑦

(b) Find the particular solution of the given d.e.(s)

(a) 2𝑥 + 3 𝑦 ' = 𝑦 + 2𝑥 + 3 1 1
2 @ (-1,0) Ans. 2𝑦 = 2𝑥 + 3 ln 2𝑥 + 3
2

(b) 𝑦 ' = 2(2𝑥 − 𝑦) @ (0,-1) Ans. 𝑦 = 2𝑥 − 1


Exercise 5.1

(a) Find the general solution of the given d.e.(s)


𝑑𝑦
(1)
𝑑𝑥 − 𝑚𝑦 = 𝐶$𝑒𝑚𝑥 , where 𝐶$ and 𝑚 are constants Ans. 𝑦 = 𝑐$𝑥 + 𝑐& 𝑒𝑚𝑥
(2) 𝑑𝑥 − 1 + 2𝑥𝑡𝑎𝑛𝑦 𝑑𝑦 = 0 Ans. 2𝑥 𝑐𝑜𝑠&𝑦 = 𝑦 + 𝑐 + 𝑠𝑖𝑛𝑦 𝑐𝑜𝑠𝑦

(b) Find the particular solution of the given d.e.(s)

(a) 2𝑥 + 3 𝑦 ' = 𝑦 + 2𝑥 + 3 1 1
2 @ (-1,0) Ans. 2𝑦 = 2𝑥 + 3 ln 2𝑥 + 3
2

(b) 𝑦 ' = 2(2𝑥 − 𝑦) @ (0,-1) Ans. 𝑦 = 2𝑥 − 1



5.2 Determination of Integrating factors
Given a d.e. of the form 𝑀 𝑑𝑥 + 𝑁 𝑑𝑦 = 0 − − 1 and a function 𝑢 = 𝑓 𝑥, 𝑦 − (2) is
to be the integrating factor of (1), then 𝑢𝑀 𝑑𝑥 + 𝑢𝑁 𝑑𝑦 = 0 − − 3 is exact, then

*(𝑢𝑀
) *(𝑢𝑁) hence 𝑢 must satisfy the partial d.e.
= *𝑥
*𝑦

𝜕𝑀 𝜕𝑢 𝜕𝑁 𝜕𝑢
𝑢 +𝑀 =𝑢 +𝑁 − −(4)
𝜕𝑦 𝜕𝑦 𝜕𝑥 𝜕𝑥
*𝑀 *𝑁
Rearranging (4) , will result to 𝑢 −
*𝑢 *𝑢 --(5)
*𝑦 *𝑥 = 𝑁*𝑥 − M *𝑦

By reversing the argument above, if 𝑢 satisfies (5), then 𝑢 is an integrating factor of (1).
No method is developed to solve (5)

Now, if we let 𝑢 = 𝑓 𝑥 , then , *𝑢 *𝑢


*𝑦 = 0 and becomes 𝑑𝑢
*𝑥 𝑑𝑥

(5) is reduced to

&
𝑢
*𝑀

*𝑁 𝑑𝑢 or *𝑀 *𝑁 𝑑𝑥 = 𝑑𝑢 − −(6) , now if
*𝑦 * = 𝑁 𝑑𝑥 𝑁 *𝑦 − *𝑥 𝑢
𝑥

& *𝑀 *𝑁
− = 𝑓 𝑥 , then, the integrating factor for (1) is 𝑢 = 𝑒 % ∫ / 𝑥 𝑑𝑥
, in the
𝑁 *𝑦 *𝑥
same argument , if 𝑢 = f y , then

& *𝑀 *𝑁 = 𝑔 𝑦 , then the integrating factor for (1) is 𝑢 = 𝑒 % ∫ 0 𝑦 𝑑𝑦

𝑀 *𝑦 − *𝑥
In summ◤ary:
1
(a) If
𝑁 #𝑀 #𝑁 = 𝑓 𝑥 , then 𝑢 = 𝑒 ' ∫ ) 𝑥 𝑑𝑥
is the integrating
#𝑦 − #𝑥
factor of
𝑀 𝑑𝑥 + 𝑁 𝑑𝑦 = 0 − −(1)
1 #𝑀
(a) (b) If #𝑁 = 𝑔 𝑦 , then 𝑢 = 𝑒 ' ∫ + 𝑦 𝑑𝑦
is the integrating
𝑀 #𝑦 − #𝑥
factor of
𝑀 𝑑𝑥 + 𝑁 𝑑𝑦 = 0 − −(1)
Example 5.2 Solve the following d.e.

(a) 𝑦 8𝑥 −◤9𝑦 𝑑𝑥 + 2𝑥 𝑥 − 3𝑦 𝑑𝑦 = 0

(b) Solve (a) by another method

(c) 𝑦 2𝑥2 − 𝑥𝑦 + 1 𝑑𝑥 + 𝑥 − 𝑦 𝑑𝑦 = 0

Solutions:

(a) Given 𝑦 8𝑥 − 9𝑦 𝑑𝑥 + 2𝑥 𝑥 − 3𝑦 𝑑𝑦 = 0 − −(1), Simplifying (1) into the form 𝑀 𝑑𝑥 +


𝑁 𝑑𝑦 = 0 will results to

8𝑦𝑥 − 9𝑦2 𝑑𝑥 + 2𝑥2 − 6𝑥𝑦 dy = 0

where : 𝑀 = 8𝑦𝑥 − 9𝑦2 and 𝑁 = 2𝑥2 − 6𝑥𝑦 , *𝑀


*𝑦 = 8𝑥 − 18𝑦 while *𝑁 = 4𝑥 − 6𝑦 , from the
*𝑥
&
form *𝑀 *𝑁 = 𝑓 𝑥 , Substituting N, *𝑀 and *𝑁
, will result to 𝑓 𝑥 = 2
, (verify this!) ,
𝑁 *𝑦 − *𝑥 *𝑦 *𝑥 𝑥

Hence , the integrating factor 𝑢 = 𝑒 ∫ / 𝑥 𝑑𝑥


= 𝑥 2 (verify this!)
Multiplying (1) by the integrating factor 𝑢 will yield

8𝑥1𝑦 − 9𝑥 2𝑦 2 𝑑𝑥 + 2𝑥2 − 6𝑥1𝑦 dy = 0 − −(2)

Now, (2) is an exact differential equation where *𝑀


*𝑦 *𝑁 = 8𝑥1 − 18𝑥2𝑦 where
= *𝑥

𝑀 = 8𝑥1𝑦 − 9𝑥 2𝑦 2 and N = 2𝑥2 − 6𝑥1𝑦 for exact d.e. such as (2) the solution is 𝐹 = 𝑐

while *𝐹
*𝑥 = 𝑀 = 8𝑥1𝑦 − 9𝑥 2𝑦 2 − −(3) and *𝐹
= 𝑁 = 2𝑥2 − 6𝑥1𝑦 − −(4)
*𝑦

Rewriting (3) into 𝜕𝐹 = (8𝑥1𝑦 − 9𝑥 2𝑦 2) 𝜕𝑥 and integrating wrt to 𝑥, and treating 𝑦 as


constant will result to 𝐹 = 2𝑥2𝑦 − 3𝑥 1𝑦 2 + 𝑇 𝑥 − − 5 , using (5) to satisfy (4)

*
𝐹 = 2𝑥2 − 6𝑥1𝑦 + 𝑇 4 𝑥 = 2𝑥2 − 6𝑥1𝑦 − −(6) , solving for 𝑇 4 𝑥 in (6) , 𝑇 4 𝑥 = 0, hence
*𝑦
𝑇 𝑥 = 𝑐 , hence the solution of (1) is 2𝑥2𝑦 − 3𝑥 1𝑦 2 = 𝑐 or 𝑥 1𝑦 2𝑥 − 3𝑦 = 𝑐
(b) Solve 𝑦 8𝑥 − 9𝑦 𝑑𝑥 + 2𝑥 𝑥 − 3𝑦 𝑑𝑦 = 0 using other method.

Solutions: Let the given d.e. 𝑦 8𝑥 − 9𝑦 𝑑𝑥 + 2𝑥 𝑥 − 3𝑦 𝑑𝑦 = 0 − − 1 be equation (1)
Now, (1) is simplified to 8𝑦𝑥 − 9𝑦2 𝑑𝑥 + 2𝑥2 − 6𝑥𝑦 dy = 0 − − 2 and (2) is
expressed in the form 𝑑𝑦
𝑑𝑥 = 𝐹(𝑥, 𝑦) which resulted to (3)

𝑑𝑦
*𝑦2$+𝑥𝑦
𝑑𝑥 = − − 3 , now (3) is to be tested for homogeneity
2𝑥2$6𝑥𝑦
2
*𝑦 $+𝑥𝑦 2
Let 𝑓 𝑥, 𝑦 = , hence 𝑓 𝑘𝑥, 𝑘𝑦 = 5𝑦
%6𝑥𝑦
= 𝑘8𝑓(𝑥, 𝑦), therefore the given d.e
2𝑥2 $6𝑥𝑦 2𝑥 2%7𝑥𝑦
is a homogenous d.e. of degree “0”.

Let 𝑦 = 𝑣𝑥 and 𝑑𝑦 = 𝑣𝑑𝑥 + 𝑥𝑑𝑣 , substituting 𝑦 and 𝑑𝑦 into (2) and simplify will result to

5𝑥2 2𝑣 − 3𝑣2 𝑑𝑥 = 2𝑥1 3𝑣 − 1 𝑑𝑣 which is variable separable which when separated


will yield 9𝑑𝑥 9𝑑𝑥
2𝑥 1𝑣%& 𝑑𝑣 which can be integrated ∫ 1𝑣%& 𝑑𝑣 , the left side is readily
= 𝑣(2%1𝑣) 2𝑥 =∫ 𝑣(2%1𝑣)
integrable but the right side is to be integrated using partial fraction.

5𝑑𝑥
Hence , the result will be ∫ ∫ '1/2𝑑 + 0/2𝑑𝑣
)⟶ 4 (𝑣𝑒𝑟𝑖𝑓𝑦 𝑡ℎ𝑖𝑠)
2𝑥 = ( 𝑣 (2'0𝑣)

Integrating (4) , the result is 5 𝑙𝑛𝑥 = − 1 𝑙𝑛𝑣 − 0 ln 2 − 3𝑣 − −(5)


2 2 2

Simplifying (5) and substituting 𝑣 = 𝑦 will yield 𝑥 0 𝑦 2𝑥 − 3𝑦 = 𝑐


𝑥
which is consistent with the answer in (a).

Note: which method do you prefer and why? Can you solve this in any other
method?
(c) Given the d.e. 𝑦 2𝑥2 − 𝑥𝑦 + 1 𝑑𝑥 + 𝑥 − 𝑦 𝑑𝑦 = 0 − −(1)

Simplifying (1) will result to 2𝑥2𝑦 − 𝑥𝑦 2 + 𝑦 𝑑𝑥 + 𝑥 − 𝑦 𝑑𝑦 = 0 − −(2)

where 𝑀 = 2𝑥2𝑦 − 𝑥𝑦 2 + 𝑦 and N = 𝑥 − 𝑦

*𝑀
Now
= 2𝑥2 − 2𝑥𝑦 + 1 and *𝑁
= 1 , it follows that &
*𝑀 *𝑁 = 2𝑥 (𝑣𝑒𝑟𝑖𝑓𝑦 𝑡ℎ𝑖𝑠)
*𝑦 *𝑥 𝑁 *𝑦 − *𝑥

Hence the integrating factor is 𝑒 ∫ 2𝑥𝑑𝑥 = 𝑒 𝑥 2 , now multiplying (2) by the integrating factor
𝑒 𝑥 2 2𝑥2𝑦 − 𝑦 2 + 𝑦 𝑑𝑥 + 𝑒 𝑥 2 𝑥 − 𝑦 𝑑𝑦 = 0 − −(3)

Now, the new 𝑀 and N will be M = 𝑒 𝑥 2 2𝑥2𝑦 − 𝑥𝑦 2 + 𝑦 and 𝑁 = 𝑒 𝑥 2 𝑥 − 𝑦

In effect of (3) becomes an exact d.e. where *𝑀


*𝑦 *𝑁 = 𝑒 𝑥 2 2𝑥2 − 2𝑥 + 1 (𝑣𝑒𝑟𝑖𝑓𝑦 𝑡ℎ𝑖𝑠)
= *𝑥
Solving (3) using the method used in exact d.e.

Let the solution of (3) be 𝐹 = 𝑐 , hence we have to find 𝐹


Also *𝐹 = N = 𝑒 𝑥2 𝑥 − 𝑦 − − 4 , we chose this instead of *𝐹
= M = 𝑒 𝑥 2 2𝑥2𝑦 − 𝑦 2 + 𝑦
*𝑦 ◤ *𝑥
why?

Now, rewriting (4) in this form 𝜕𝐹 = 𝑒 𝑥 2 𝑥 − 𝑦 𝜕𝑦 and integrating (4) wrt 𝑦, holding 𝑥
2
constant as shown. ∫ 𝜕𝐹 = ∫ 𝑒𝑥 𝑥 − 𝑦 𝜕𝑦 will results to

𝐹 = 𝑒 𝑥 2 𝑥𝑦 − & 𝑦 2 + 𝑇 𝑥 − − 5 , What is the significance of 𝑇 𝑥 in (5)?


2

Also , since *𝐹
*𝑥 = M , finding *𝐹 in (5), differentiating wrt 𝑥 while holding 𝑦 constant will
*𝑥
results to 𝑒 𝑥 2 𝑦 + 2𝑥2 − 𝑥𝑦 2 + 𝑇 4 𝑥 − − 6 (𝑣𝑒𝑟𝑖𝑓𝑦 𝑡ℎ𝑖𝑠)

Equating (6) to M, 𝑒 𝑥 2 𝑦 + 2𝑥2𝑦 − 𝑥𝑦 2 + 𝑇 4 𝑥 = 𝑒 𝑥 2 2𝑥2𝑦 − 𝑥𝑦2 + 𝑦 and finding 𝑇 4 𝑥 ,

𝑇 4 𝑥 = 0, which when integrated will give 𝑇 𝑥 = 𝑐 , hence the general solution of (1) is
𝑒 𝑥 2 𝑦 2𝑥 − 𝑦 = 𝑐𝑒%𝑥2 (𝑣𝑒𝑟𝑖𝑓𝑦 𝑡ℎ𝑖𝑠)
Exercise◤s 5.2 Solve each of the following d.e.

(a) 𝑥𝑦 + 1 𝑑𝑥 + 𝑥 𝑥 + 4𝑦 − 2 𝑑𝑦 = 0

Ans. 𝑥𝑦 + 𝑙𝑛 𝑥 + 2𝑦R − 2𝑦 = 𝑐
(a) 𝑦 R𝑑𝑥 + 3𝑥𝑦 + 𝑦 R + 1 𝑑𝑦 = 0 Ans. 𝑦 2𝑥 + 𝑦 = 𝑐𝑒 𝑥

(b)2 2𝑦R + 5𝑥𝑦 − 2𝑦 + 4 𝑑𝑥 + 𝑥 2𝑥 + 2𝑦 − 1 𝑑𝑦 = 0

Ans. 𝑥 T 𝑦 R + 2𝑥𝑦 − 𝑦 + 2 = 𝑐
(a) 3 𝑥 R + 𝑦 R 𝑑𝑥 + 𝑥 𝑥 R + 3𝑦R + 6𝑦 𝑑𝑦 = 0.

Ans. 𝑥 𝑥 R + 3𝑦R = 𝑐𝑒 –𝑦

You might also like