Ch8 PDF
Ch8 PDF
2016年9月19日 下午 06:33
微分方程式
Chapter 8
管傑雄
台大電機系
Chapter 8 Systems of Linear First-order
Differential Equation
8.1 Preliminary Theory
Note
System of differential equations
Ch8 第 1 頁
..............
where aij and fi are functions continuous on a common interval I. When fi(t) = 0, i = 1, 2,
3, ... , n, the system is said to be homogeneous; otherwise it is called nonhomogeneous. We
also refer to the above system as nth-order linear system.
Note:
1. Equation to a System
Suppose a linear nth-order differential equation is first written as
3. Degenerate Systems
Those systems of differential equations cannot be reduced to a linear system in normal
form i.e., we cannot represent the highest-order terms with the others, are said to be
degenerate.
Example: Reduce
(D + 1)x + (D + 1)y = 0
2Dx + (2D + 1)y = 0
to the normal form.
Ch8 第 2 頁
to the normal form.
Solution: Actually the system is degenerate since the determinant of the coefficients of the
highest-order terms is zero and we can not.
Matrix Functions
is called a matrix function or a matrix-valued function. If each aij(t) is continuous, then A(t) is
said to be continuous.
Note:
1. Derivative of a Matrix Function
, C is a matrix of constants
where
Ch8 第 3 頁
Initial-Value Problem
Solve:
Superposition Principle
Let X1, X2, ... , Xk be a set of solution vectors of the homogeneous system X' = AX on
an interval I. Then the linear combination
X = c1X1 + c2X2 + ... + ckXk
where ci, i = 1, 2, ... , k, are arbitrary constants, is also a solution in the interval.
Linear Dependence/Independence
Let X1, X2, ... , Xk be a set of solution vectors of the homogeneous system X' = AX on
an interval I. We say that the set is linearly dependent on the interval, if there exist
contants c1, c2, ... , ck, not all zero, such that
c1X1 + c2X2 + ... + ckXk= 0
for every t in the interval. If the set of vectors is not linearly dependent on the interval, it is
said to be linearly independent.
, ........
for every t in I.
Ch8 第 4 頁
for every t in I.
In fact, it can be shown that if X1, X2, ... , Xn are solution vectors of X' = AX, then for
every t in I either
or
Thus, if we can show that W 0 for some t0 in I, then W 0 for every t and hence the
solutions are linearly independent on the interval.
Proof: We try to confirm the following statements: If there is a specific value t0 in the
interval to make W = 0, then W = 0 for every t in the interval.
From the assumption, we find that X1(t0), X2(t0), ... , Xn(t0) are linearly dependent. This
indicates that we can find n constants c1, c2, ... , cn, not all zero to make
X0 = c1X1(t0) + c2X2(t0) + ... + cnXn(t0) = 0
Using these constant, we can form a vector
X(t) = c1X1(t) + c2X2(t) + ... + cnXn(t)
In particular, X(t0) = 0. From the superposition principle, we know that X(t) is a solution of X'
= AX. If we solve the differential equation X' = AX subject to X(t0) = X0 = 0, the solution
vector should be unique, and we find that it is the zero vector. That is, X(t) = 0. for every t in
the interval. Since ci is not all zero, X1(t), X2(t), ... , Xn(t) are linearly dependent for every t in
the interval. Therefore, W = 0 for every t in the interval.
, ... ,
,
Since W(X1(t0), X2(t0), ... , Xn(t0)) 0, we conclude that
for every t and they are linearly independent. Therefore, they are a fundamental set of
solutions.
Example: Solve
and then
and so eigenvalues are -3, -4, and 5. The associated eigenvectors are
Ch8 第 6 頁
,
Example: Solve
and
Ch8 第 7 頁
(2) If there is only one eigenvector corresponding to the eigenvalue of multiplicity m, then m
linearly independent solutions of the form
........
Eigenvalue of multiplicity 2
We try the solution form
Hence,
(A - 1I)K = 0
(A - 1I)P = K
We can find K first and then find P.
Example: Solve
P can be gievn by
.
The two independent solutions are
Ch8 第 8 頁
Eigenvalue of Multiplicity 3
We try the solution form
Example: Solve
and
is a solution.
Ch8 第 9 頁
is a solution.
改寫成Eigenvector的實部及虛部
By defining
and
then
Example: Solve
The matrices
, and
Ch8 第 10 頁
,
, ........
is said to be a fundamental matrix of the system on the interval. Then the general solution of
the homogeneous system X' = AX can be written as
X = C
where C is an
column vector of arbitrary constants.
Note:
1. The Fundamental Matrix Has an Inverse
Let (t) be a fundamental matrix of the homogeneous system X' = AX on an interval I.
Then -1(t) exists for every value of t in the interval.
2. Special Matrix
Let V1, V2, ... , Vn be solution vectors of X' = AX that satisfy the conditions
Ch8 第 11 頁
, ... ,
Here t0 is an arbitrary chosen point in the interval on which the general solution of the system
is defined. The matrix
=I
where I is the
multiplicative identity.
Note: Suppose a constant matrix
such that
i.e.,
and
Example:
and
for every t in the I. They form a fundamental set of solutions and the general solution is
X = c1X1 + c2X2 = C
where
Ch8 第 12 頁
and
The special fundamental matrix (t) satisfying (0) = I can be found from the column
vectors:
and
and
and
Example: solve
Ch8 第 13 頁
and the roots are -2, 1 and 5. The associated eigenvectors are
The matrix D is
The solution of X is
Ch8 第 14 頁
.
The complementary function is
and no coincidence of the terms with the complementary function occurs. Then we can
assume the particular function is
Note: If there is a coincidence of terms, the same rule as the nth-order case should be
followed.
2. Variation of Parameters
The nonhomogeneous system
X' = AX + F(t)
can solved by first solving the homogeneous system
X' = AX
The associated general solution, i.e., the complementary function, is
Xc = (t)C
where (t) is a fundamental matrix of the system. By using variation of parameters to find
the particular solution, we can replace the constant vector C with a column matrix of
functions
Xp = (t)U(t)
where
Note: 上述式子在計算矩陣積分時須特別注意:對矩陣各element的積分,積分下限必須
一致。
Example: Solve
Ch8 第 15 頁
Solution: The characteristic equation of the coefficient matrix is
( + 2)( + 5) = 0
The roots are -2 and -5. The associated eigenvectors are
.
The fundamental matrix is
In the nonhomogeneous system, let the initial condition be X = X0 at t = t0. Since the
general solution of the nonhomogeneous system can be written as
.
The general solution can be rewritten as
Note: Diagonalization
If P is the matrix that diagonalizes the coefficient matrix A, and X is transformed with
X = PY, the differential equation of Y is
Y' = P-1APY + P-1F or Y' = DY + G
where D = P-1AP and G = P-1F.
Example: Solve
by diagonalization.
Solution: The characteristic equation of A is
2 - 5 = 0
Ch8 第 16 頁
2 - 5 = 0
The roots are 0 and 5 and the associated eigenvectors are
The matrix G is
求解常係數聯立微分方程組的程序步驟
1. 降階方程組成為Normal Form X’ = AX + F(t),其中A是常係數矩陣
2. 求出A的Eigenvalue及Eigenvector。對於重根時,若有必要則須找出K, P, Q….
3. 由2之結果求出足夠的獨立解,針對共軛解則改以實部及虛部代之
4. 將獨立排成Fundamental Matrix
5. 計算Special Fundamental Matrix
.
6. 計算通解
Note: 1.
2.
Ch8 第 17 頁
3. etAe-sA = e(t-s)A
4. etA is a fundamental matrix of the system X' = AX: (t) = etA
The result of 2. indicates that the solution of X' = AX + F(t) can be written as
.
Solution: The matrix A = I + B where I is the identity matrix while
where
and
since
Note: The simplicity of the computations in this example is due to the fact that A is an upper
triangular matrix.
Ch8 第 18 頁
2016.12.07 (三) OK
2016年12月7日 下午 03:29
Ch8 第 19 頁
Ch4方法 system of DE's
2016年12月7日 下午 03:29
考慮聯立方程式
其中x最高m階;y最高n階
可將其寫成
寫成矩陣形式
Case 1:
x, y 具有唯一解
先求出齊性解
、
須留意此處變成最高(n+m)階微分,且x與y齊性解相同,但特解不同,可得
再求出特解
代回原微分方程式任意一條,可以解得 關係,最後待定常數c的個數為(n+m)個
Case 2:
且 且 則有無限多組解
Case 3:
且 或 則無解
Ch8 第 20 頁
Example
2016年12月7日 下午 03:56
代回原式,解得
待定係數僅剩 三個
Ch8 第 21 頁
Ch8 方法
2016年12月7日 下午 03:56
Step 1: Normalization
若 ,可以將x, y的最高階微分項分離
(初步normalization)
若 ,則為degenerate,可以將方程式階數降低
然後繼續進行normalization
Example : (degenerate)
(降階)
代回原方程式
Ch8 第 22 頁
可表示為
此為將n-th order linear DE轉成system of linear equation
將初步normalization 的結果改寫成
寫成矩陣型式
Ch8 第 23 頁
可簡化表示為
此為將聯立 x:m-th order & y:n-th order linear DE轉成system of linear equation
Ch8 第 24 頁
2016.12.09 (五) ok
2017年1月24日 上午 11:17
Ch8 第 25 頁
矩陣解 normal form
2017年1月24日 上午 11:18
考慮
其中A為 矩陣
Initial condition
僅考慮唯一解析解區之解
首先求解Homogeneous 情況
求得
令
Solution set :
Fundamental matrix :
若 則 有反矩陣
再考慮Nonhomogeneous
求出一個
方程式之解為
Ch8 第 26 頁
Case 1 相異實根
2017年1月24日 上午 11:19
猜
其中K為constant colume vector
代入方程式可得
.…..characteristic equation
上式為n次方程式,依高斯代數基本定理 可以解出n個解
若 為相異實根,則
Example: Solve
and then
and so eigenvalues are -3, -4, and 5. The associated eigenvectors are
Ch8 第 27 頁
The general solution is
Ch8 第 28 頁
Case 2 重根
2017年1月24日 上午 11:19
Example: Solve
and
(2) If there is only one eigenvector corresponding to the eigenvalue of multiplicity m, then m
linearly independent solutions of the form
........
Eigenvalue of multiplicity 2
We try the solution form
Ch8 第 29 頁
We try the solution form
Hence,
(A - 1I)K = 0
(A - 1I)P = K
We can find K first and then find P.
Example: Solve
P can be gievn by
.
The two independent solutions are
Eigenvalue of Multiplicity 3
We try the solution form
Example: Solve
Ch8 第 30 頁
(A - 1I)K = 0, the solution is
*觀念統整
並非雙重根就一定要找K、P,只要找到兩個eigen vectors 即可;並非三重根就一定
要找K、P、Q,只要找到eigen vectors即可。
對於三重根而言,有以下幾個可能性
1.
2.
3.
4.
但需要留意 1. 的向量空間只到 ;2. 與 3. 的向量空
間有 與 ;4. 的向量空間有 、 與 。
對於一組微分方程,雖然可以任意選擇組合,但是不可能改變解的向量空間。
*由K、P、Q反求A矩陣 (自己出考題)
設定 三重根
Ch8 第 31 頁
設定 三重根
給定eigen vector
Ch8 第 32 頁
Case 3 複數根
2017年1月24日 上午 11:19
Let A be the coefficient matrix having real entries of X' = AX and let K1 be an eigenvector
corresponding to the complex eigenvalue 1 = + i, and real. Then
X1 = K1e t and
1
and
is a solution.
改寫成Eigenvector的實部及虛部
By defining
and
then
Example: Solve
Ch8 第 33 頁
The matrices
, and
Ch8 第 34 頁
2016.12.16 (五) ok
2017年1月24日 上午 11:20
Ch8 第 35 頁
Solution of X'=AX
2017年1月24日 上午 11:22
可寫為
Ch8 第 36 頁
求解 Xp
2017年1月24日 上午 11:22
Ch8 第 37 頁
Undetermined method
2017年1月24日 上午 11:22
Ch8 第 38 頁
Variation of parameter
2017年1月24日 上午 11:22
Example: Solve
Ch8 第 39 頁
( + 2)( + 5) = 0
The roots are -2 and -5. The associated eigenvectors are
.
The fundamental matrix is
Ch8 第 40 頁
Diagonalization
2017年1月24日 上午 11:23
代入
其中 為對角線化矩陣(非微分算符),
可以得知 、
容易解Y,解出後將Y代回
Ch8 第 41 頁
求解常係數聯立微分方程組的程序步驟
2017年1月26日 下午 04:46
.
6. 計算通解
或者是
Ch8 第 42 頁
2016.12.21 (三) ok
2017年1月24日 上午 11:23
Ch8 第 43 頁
Matrix function
2017年1月24日 上午 11:23
考慮一個函數 ,假設可以使用泰勒展開
定義Matrix function
A為n*n方陣。可以注意到,使用power series定義矩陣函數可以保證其存在,並且函
數值也是矩陣。
現在考慮一個函數
其對應的矩陣函數為
以下為矩陣函數 之性質。
1.定義 (A為零矩陣也成立)
因此
2.
3.
滿足反矩陣定義,因此 與 互為反矩陣
4.
Proof:
Ch8 第 44 頁
特別注意到 滿足
這提供了簡便的方法來求
用此公式容易寫成程式
.
Solution: The matrix A = I + B where I is the identity matrix while
where
and
since
Note: The simplicity of the computations in this example is due to the fact that A is an upper
triangular matrix.
Ch8 第 45 頁
Ch8 第 46 頁
Ch8 第 47 頁