没有合适的资源?快使用搜索试试~ 我知道了~
首页透彻理解微分方程:线性与非线性、常微分与偏微分解题指南
透彻理解微分方程:线性与非线性、常微分与偏微分解题指南
需积分: 9 19 下载量 91 浏览量
更新于2023-07-22
1
收藏 6.76MB PDF 举报
《线性、非线性、常微分方程与偏微分方程》是一本专为科学、工程和数学专业人士精心编撰的教材,旨在帮助读者深入理解和掌握微分方程的核心概念。该书由A.C. King、J.Billingham和S.R.Otto三位作者共同编写,特别强调构建解析解的方法,并通过MATLAB软件进行大量实例演示,这些实例源于现实生活中的有趣且不寻常的问题,以增强理论与实践的结合。 书中内容涵盖了四个主要的微分方程类型:线性方程,它们通常具有明确的结构,可以利用诸如特征值和特征向量等工具求解;非线性方程,则更复杂,可能没有封闭形式的解,但可通过数值方法或变分法等手段求近似解;常微分方程主要研究的是只有一个时间变量的系统,例如牛顿运动定律的应用;而偏微分方程则涉及到多个空间变量,广泛应用于物理学、工程学等领域,如热传导、波动和电磁学问题。 作者不仅注重理论教学,还提供了大量的练习题,让读者能够通过实际操作巩固所学知识。此外,书中强调理解解的物理意义,这在实际问题解决中至关重要。书末尾的版权信息提到,尽管本书享有版权,但在法律允许的范围内和集体许可协议的框架下,读者可以复制部分内容,但必须事先获得剑桥大学出版社的书面许可。 《线性、非线性、常微分方程与偏微分方程》是学习和研究微分方程理论和应用的宝贵资源,对于希望在这个领域深化理解的学生和研究人员来说,它既是一本实用的学习指南,也是一本探索数学之美和科学奥秘的参考书。通过阅读和实践书中的内容,读者将能够熟练掌握微分方程的分析技巧,以及如何将其应用于解决实际问题。
资源详情
资源推荐
![](https://csdnimg.cn/release/download_crawler_static/830319/bg10.jpg)
1.2 THE METHOD OF VARIATION OF PARAMETERS 9
1.2.1 The Wronskian
Before we carry on, let’s pause to discuss some further properties of the Wronskian.
Recall that if V is a vector space over R, then two elements v
1
, v
2
∈ V are linearly
dependent if ∃ α
1
,α
2
∈ R, with α
1
and α
2
not both zero, such that α
1
v
1
+α
2
v
2
= 0.
Now let V = C
1
(a, b) be the set of once-differentiable functions over the interval
a<x<b.Ifu
1
,u
2
∈ C
1
(a, b) are linearly dependent, ∃ α
1
,α
2
∈ R such that
α
1
u
1
(x)+α
2
u
2
(x)=0∀x ∈ (a, b). Notice that, by direct differentiation, this also
gives α
1
u
1
(x)+α
2
u
2
(x) = 0 or, in matrix form,
u
1
(x) u
2
(x)
u
1
(x) u
2
(x)
α
1
α
2
=
0
0
.
These are homogeneous equations of the form
Ax = 0,
which only have nontrivial solutions if det(A) = 0, that is
W =
u
1
(x) u
2
(x)
u
1
(x) u
2
(x)
= u
1
u
2
− u
1
u
2
=0.
In other words, the Wronskian of two linearly dependent functions is identically
zero on (a, b). The contrapositive of this result is that if W ≡ 0on(a, b), then u
1
and u
2
are linearly independent on (a, b).
Example
The functions u
1
(x)=x
2
and u
2
(x)=x
3
are linearly independent on the interval
(−1, 1). To see this, note that, since u
1
(x)=x
2
, u
2
(x)=x
3
, u
1
(x)=2x, and
u
2
(x)=3x
2
, the Wronskian of these
two functions is
W =
x
2
x
3
2x 3x
2
=3x
4
− 2x
4
= x
4
.
This quantity is not identically zero, and hence x
2
and x
3
are linearly independent
on (−1, 1).
Example
The functions u
1
(x)=f(x) and u
2
(x)=kf(x), with k a constant, are linearly
dependent on any interval, since their Wronskian is
W =
fkf
f
kf
=0.
If the functions u
1
and u
2
are solutions of (1.2), we can show by differentiating
W = u
1
u
2
− u
1
u
2
directly that
dW
dx
+ a
1
(x)W =0.
![](https://csdnimg.cn/release/download_crawler_static/830319/bg11.jpg)
10 VARIABLE COEFFICIENT, SECOND ORDER DIFFERENTIAL EQUATIONS
This first order differential equation
has solution
W (x)=W (x
0
) exp
−
x
x
0
a
1
(t)dt
, (1.7)
which is known as Abel’s formula. This gives us an easy way of finding the
Wronskian of the solutions of any second order differential equation without having
to construct the solutions themselves.
Example
Consider the equation
y
+
1
x
y
+
1 −
1
x
2
y =0.
Using Abel’s formula, this has Wronskian
W (x)=W (x
0
) exp
−
x
x
0
dt
t
=
x
0
W (x
0
)
x
=
A
x
for some constant A. To find this constant, it is usually necessary to know more
about the solutions u
1
(x) and u
2
(x). We will describe a technique for doing this in
Section 1.3.
We end this section with a couple of useful theorems.
Theorem 1.1 If u
1
and u
2
are linearly independent solutions of the homoge-
neous, nonsingular ordinary differential equation (1.2), then the Wronskian is either
strictly positive or strictly negative.
Proof From Abel’s formula, and since the exponential function does not change
sign, the Wronskian is identically positive, identically negative or identically zero.
We just need to exclude the possibility that W is ever zero. Suppose that W (x
1
)=
0. The vectors
u
1
(x
1
)
u
1
(x
1
)
and
u
2
(x
1
)
u
2
(x
1
)
are then linearly dependent, and
hence u
1
(x
1
)=ku
2
(x
1
) and u
1
(x)=ku
2
(x) for some constant k. The function
u(x)=u
1
(x) −ku
2
(x) is also a solution of (1.2) by linearity, and satisfies the initial
conditions u(x
1
)=0,u
(x
1
) = 0. Since (1.2) has a unique solution, the obvious
solution, u ≡ 0, is the only solution. This means that u
1
≡ ku
2
. Hence u
1
and u
2
are linearly dependent – a contradiction.
The nonsingularity of the differential equation is crucial here. If we consider the
equation x
2
y
− 2xy
+2y = 0, which has u
1
(x)=x
2
and u
2
(x)=x as its linearly
independent solutions, the Wronksian is −x
2
, which vanishes at x = 0. This is
because the coefficient of y
also vanishes at x =0.
Theorem 1.2 (The Sturm separation theorem) If u
1
(x) and u
2
(x) are the
linearly independent solutions of a nonsingular, homogeneous equation, (1.2), then
![](https://csdnimg.cn/release/download_crawler_static/830319/bg12.jpg)
1.3 SOLUTION BY POWER SERIES: THE METHOD OF FROBENIUS 11
the zeros of u
1
(x) and u
2
(x) occur alternately. In other words,
successive zeros of
u
1
(x) are separated by successive zeros of u
2
(x) and vice versa.
Proof Suppose that x
1
and x
2
are successive zeros of u
2
(x), so that W (x
i
)=
u
1
(x
i
)u
2
(x
i
) for i = 1 or 2. We also know that W (x) is of one sign on [x
1
,x
2
],
since u
1
(x) and u
2
(x) are linearly independent. This means that u
1
(x
i
) and u
2
(x
i
)
are nonzero. Now if u
2
(x
1
) is positive then u
2
(x
2
) is negative (or vice versa), since
u
2
(x
2
) is zero. Since the Wronskian cannot change sign between x
1
and x
2
, u
1
(x)
must change sign, and hence u
1
has a zero in [x
1
,x
2
], as we claimed.
As an example of this, consider the equation y
+ω
2
y = 0, which has solution y =
A sin ωx + B cos ωx. If we consider any two of the zeros of sin ωx, it is immediately
clear that cos ωx has a zero between them.
1.3 Solution by Power Series: The Method of Frobenius
Up to this point, we have considered ordinary differential equations for which we
know at least one solution of the homogeneous problem. F
rom this we have seen that
we can easily construct the second
independent solution and, in the inhomogeneous
case, the particular in
tegral. We now turn our attention to the more difficult
case, in which we cannot determine a solution of the homogeneous problem by
inspection. We must devise a method that is capable of solving variable coefficient
ordinary differential equations in general. As we noted at the start of the chapter,
we will restrict our attention to the case where the variable coefficients are simple
polynomials. This suggests that we can look for a solution of the form
y = x
c
∞
n=0
a
n
x
n
=
∞
n=0
a
n
x
n+c
, (1.8)
and hence
dy
dx
=
∞
n=0
a
n
(n + c)x
n+c−1
, (1.9)
d
2
y
dx
2
=
∞
n=0
a
n
(n + c)(n + c − 1)x
n+c−2
, (1.10)
where the constants c, a
0
,a
1
,... , are as yet undetermined. This is known as the
method of Frobenius. Later on, we will give some idea of why and when this
metho
d can be used. For the moment, we will just try to make it work. We proceed
by example, with the simplest case first.
1.3.1 The Roots of the Indicial Equation Differ by an Integer
Consider the equation
x
2
d
2
y
dx
2
+ x
dy
dx
+
x
2
−
1
4
y =0. (1.11)
![](https://csdnimg.cn/release/download_crawler_static/830319/bg13.jpg)
12 VARIABLE COEFFICIENT, SECOND ORDER DIFFERENTIAL EQUATIONS
We substitute (1.8) to (1.10) in
to (1.11), which gives
x
2
∞
n=0
a
n
(n + c)(n + c − 1)x
n+c−2
+ x
∞
n=0
a
n
(n + c)x
n+c−1
+
x
2
−
1
4
∞
n=0
a
n
x
n+c
=0.
We can rearrange this slightly to obtain
∞
n=0
a
n
(n + c)(n + c − 1)+(n + c) −
1
4
x
n+c
+
∞
n=0
a
n
x
n+c+2
=0,
and hence, after simplifying the terms in the first summation,
∞
n=0
a
n
(n + c)
2
−
1
4
x
n+c
+
∞
n=0
a
n
x
n+c+2
=0.
We now extract the first two terms from the first summation to give
a
0
c
2
−
1
4
x
c
+ a
1
(c +1)
2
−
1
4
x
c+1
+
∞
n=2
a
n
(n + c)
2
−
1
4
x
n+c
+
∞
n=0
a
n
x
n+c+2
=0. (1.12)
Notice that the first term is the only one containing x
c
and similarly for the second
term in x
c+1
.
The two summations in (1.12) begin at the same power of x, namely x
2+c
.If
we
let m = n + 2 in the last summation
(notice that if
n = 0 then m = 2, and n = ∞
implies that m = ∞), (1.12) becomes
a
0
c
2
−
1
4
x
c
+ a
1
(c +1)
2
−
1
4
x
c+1
+
∞
n=2
a
n
(n + c)
2
−
1
4
x
n+c
+
∞
m=2
a
m−2
x
m+c
=0.
Since the variables in the summations are merely dummy variables,
∞
m=2
a
m−2
x
m+c
=
∞
n=2
a
n−2
x
n+c
,
and hence
a
0
c
2
−
1
4
x
c
+ a
1
(c +1)
2
−
1
4
x
c+1
+
∞
n=2
a
n
(n + c)
2
−
1
4
x
n+c
+
∞
n=2
a
n−2
x
n+c
=0.
![](https://csdnimg.cn/release/download_crawler_static/830319/bg14.jpg)
1.3 SOLUTION BY POWER SERIES: THE METHOD OF FROBENIUS 13
Since the last two summations in
volve identical powers of
x, we can combine them
to obtain
a
0
c
2
−
1
4
x
c
+ a
1
(c +1)
2
−
1
4
x
c+1
+
∞
n=2
a
n
(n + c)
2
−
1
4
+ a
n−2
x
n+c
=0. (1.13)
Although the operations above are straightforward, we need to take some care to
avoid simple slips.
Since (1.13) must hold for all values of x, the coefficient of each power of x must
be zero. The coefficient of x
c
is therefore
a
0
c
2
−
1
4
=0.
Up to this point, most Frobenius analysis is very similar. It is here that the different
structures come into play. If we were to use the solution a
0
= 0, the series (1.8)
would have a
1
x
c+1
as its first term. This is just equivalent to increasing c by 1. We
therefore assume that a
0
= 0, which means that c must satisfy c
2
−
1
4
= 0. This is
called the indicial equation, and implies that c = ±
1
2
. Now, progressing to the
next term, proportional to x
c+1
, we find that
a
1
(c +1)
2
−
1
4
=0.
Choosing c =
1
2
gives a
1
= 0, and, if we were to do this, we would find that we had
constructed a solution with one arbitrary constant. However, if we choose c = −
1
2
the indicial equation is satisfied for arbitrary values of a
1
, and a
1
will act as the
second arbitrary constant for the solution. In order to generate this more general
solution, we therefore let c = −
1
2
.
We now progress to the individual terms in the summation. The general term
yields
a
n
n −
1
2
2
−
1
4
+ a
n−2
= 0 for n =2, 3,... .
This is called a recurrence relation. We solve it by observation as follows. We
start by rearranging to give
a
n
= −
a
n−2
n(n −1)
. (1.14)
By putting n = 2 in (1.14) we obtain
a
2
= −
a
0
2 ·1
.
For n =3,
a
3
= −
a
1
3 ·2
.
剩余547页未读,继续阅读
![pdf](https://img-home.csdnimg.cn/images/20210720083512.png)
![image/x.djvu](https://img-home.csdnimg.cn/images/20210720083646.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
安全验证
文档复制为VIP权益,开通VIP直接复制
![](https://csdnimg.cn/release/wenkucmsfe/public/img/green-success.6a4acb44.png)