MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂
nnnØØØ���¢¢¢~~~
Xi Chen
ibs.bfsu.edu.cn/chenxi
Department of Management Science and Engineering
International Business School
Beijing Foreign Studies University
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 1 / 139
�µ�¦
1 p�êÆ!�5�ê!VÇض
2 $ÊÆ!ÚOƶ
3
) C �ó½a C �ó�Ä�?§(�”
¡�é�
1 �c?ïÄ)¶
2 nc?��)”
8IµU
|^ MATLAB ¥Jø��«Ä�õUÚp?óä�©Û²
L�+nÆ¥�ïįK”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 2 / 139
MATLAB ÄÄÄ:::
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 3 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 4 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
Ý
�ê|$�
Ý
Ñ\
A =
1 2 34 5 6
7 8 9
3 MATLAB ·-I�¥§Ñ\
A = [1, 2, 3; 4, 5, 6; 7, 8, 9]
Ue Enter �w«(J”
Ý
A ��±?\ MATLAB ?6ì¥Ñ\±e§S�è¢y
for i=1:1:3
for j=1:1:3
A(i,j)=(i-1)*3+j;
end
end
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 5 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
Ý
��
3 MATLAB ·-I�¥§�gÑ\e�·-
>> x=[-1.3,sqrt(3),(1+2+3)*4/5]
>> x(5)=abs(x(1))
>> x(4)=abs(x(2))
e
A =
1 2 34 5 6
7 8 9
3 MATLAB ·-I�¥§�gÑ\e�·-
>> B=[A;[10,11,12]]
>> C=A(1:2,:)
>> D=A(1:2,1:2)
>> E=A([1,3],[1,3])
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 6 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
Ý
$�
e A = [1, 2, 3; 4, 5, 6; 7, 8, 9]§* B = A′, C = A + B �(J¶
e x = [−1, 0, 2]§* y = x − 1 �(J¶
e A = [1, 2, 3; 4, 5, 6; 7, 8, 9]§* pi ∗ A, A ∗ pi �(J¶
e X = [−1, 0, 2], Y = [−2,−1, 1]§* X ∗ Y ′, X ′ ∗ Y �(J¶
¦)�àg�5�§|
2×1 + x2 − 5×3 + x4 = 8
x1 − 3×2 − 6×4 = 9
2×2 − x3 + 2×4 = −5
x1 + 4×2 − 7×3 + 6×4 = 0
�âK¿3 MATLAB ·-I�¥ïáåXêÝ
A Úmà��
þ b§* det(A), A\b, inv(A) ∗ b, b/A, b ∗ inv(A) �(J”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 7 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
Example 1
2×1 − 4×2 + 5×3 = −4
−4×1 − 2×2 + 3×3 = 4
2×1 + 6×2 − 8×3 = 0
�g$1µrank(A), rank([A,b]), null(A), pinv(A)*b
Example 2
|^”�m(nullspace, AX = 0)y²eA�÷�§KATA�_”¿¦)µ
A =
1 0 1
2 −1 1
−2 3 −2
0 1 5
b =
1
0
1
6
�g$1µrank(A), rank([A,b]), A\b, inv(A’*A)*A’*b, pinv(A)*b
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 8 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
ê|$�
\~$�¶
e x = [1, 2, 3], y = [4, 5, 6]§* x . ∗ y , x .\y , x ./y �(J¶
e x = [1, 2, 3], y = [4, 5, 6]§* x .ˆy , x .ˆ2, 2.ˆx �(J”
�þ�eI
* x = 1 : 1 : 5, x = 6 : −1 : 1 �(J¶
�ã* y = limx→0 sin x/x �4��
>> x=pi/2:-0.00001:0.00001
>> y=sin(x)./x
>> plot(x,y,’o’)
e A = [1, 2, 3; 4, 5, 6; 7, 8, 9]§
1 * A(1 : 3, 3), A(1 : 2, 2 : 3), A(:, 2) �(J¶
2 * A([1, 3], [1, 3]), A([1, 3], 🙂 �(J¶
3 |^Ð�1C�ò A {z��F/§¿� rref (A) ¼ê(J’�”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 9 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
�é�Cþ
5¿���«©§“;”Ú“…”��^¶
eps, pi , i , inf , nan þ�XÚ�3Cþ¶
‘X$�¥Iþ�Iþ§Iþ�Ý
§Ý
�Ý
�L�ª¶
e A = [1, 2, 3; 2, 3, 4], B = [1, 2, 2; 2, 2, 3]§*
x = (1 < 2), C = (A <= B), C = (A <= 1) �(J¶
Ü6$�¥Iþ�Iþ§Iþ�Ý
§Ý
�Ý
�L�ª¶
e A = [0, 2, 3; 0, 2, 0], B = [0, 0, 0; 2, 3, 4]§* A&B, A|B, ˜A �
(J¶
~^·-µformat, help, quit, save, save filename, save filename x y
z, clear, clc, who, whos, ↑¶ÁO�
xn+1 =
1
2
(
xn +
2
xn
)
, x1 = 2,
√
1 +
√
1 +
√
1 + . . .
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 10 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
Example 3
ln (1 + x) = x −
x2
2
+
x3
3
− . . .+ (−1)n−1
xn
n
+ . . . (−1 < x ≤ 1)
|^Si¼ê��O�ln 2��¶
|^�©){c20��ÚCqO�ln 2��¶
|^eªc3��ÚCqO�ln 2��"
ln
1 + x
1− x
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 11 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
Example 4
arctan x = x −
x3
3
+
x5
5
−
x7
7
+ . . .+ (−1)n−1
x2n−1
2n − 1
+ . . . (−1 ≤ x ≤ 1)
|^c20��ÚCqO�π�¶
e|^î.úª5O�π�§`²O��Ç´ÄkJp¶
π
4
= arctan
1
2
+ arctan
1
3
e|^e�_íª2O�π�§`²O��Ç´ÄkJp"
an+1 =
√
an + 1/
√
an
2
, bn+1 =
√
an(1 + bn)
an + bn
pn+1 =
pnbn+1(1 + an+1)
1 + bn+1
, a0 =
√
2, b0 = 0, p0 = 2 +
√
2
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 12 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
Iterative methods for solving a square linear system
We seek an iteration of the form xk+1 = F (xk), where an initial guess
x0 ∈ Rn is given and F is simple to compute.
Jacobi Iteration
Ax = b
(U + L + D)x = b
Dx = −(U + L)x + b
x = −D−1(U + L)x + D−1b
Gauss-Seidel Iteration
Ax = b
(U + L + D)x = b
(L + D)x = −Ux + b
x = −(L + D)−1Ux + (L + D)−1b
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 13 / 139
MATLAB ÄÄÄ::: ÝÝÝ
���êêê|||$$$���
Example 5
Using Jacobi and Guass-Seidel iteration to solve the following linear
equation system and show their computational efficiency.
7x1 + x2 + 2x3 = 10
x1 + 8x2 + 2x3 = 8
2x1 + 2x2 + 9x3 = 6
Theorem 6 (Weierstrass Approximation Theorem)
If f (x) is a continuous function in [a, b], then for every � > 0, there exists
a polynomial p(x) such that
|f (x)− p(x)| ≤ �
for every x ∈ [a, b].
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 14 / 139
MATLAB ÄÄÄ::: ~~~^̂̂êêêÆÆƼ¼¼êêê
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 15 / 139
MATLAB ÄÄÄ::: ~~~^̂̂êêêÆÆƼ¼¼êêê
~^êƼê
êâÚO
�þ�Ý
��������§max, min¶
ex = [9, 4, 7, 8, 3]§*
max(x)§min(x)§[y , i ] = max(x)§[y , i ] = min(x)�(J¶
eA = [1, 5, 3; 7, 2, 6; 9, 3, 8]§*
max(A)§max(A, [ ], 1)§max(A, [ ], 2)�(J”
¦Ú�¦È§sum, prod¶
²þ�ÚIO��§mean, std(A, flag , dim)
σ1 =
√√√√ 1
n − 1
n∑
i=1
(xi − x)2 ½ σ2 =
√√√√1
n
n∑
i=1
(xi − x)2
�’Xê§corrcoef
r =
√∑n
i=1(xi − x)(yi − y)√∑n
i=1(xi − x)2 ·
√∑n
i=1(yi − y)2
üS§sort”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 16 / 139
MATLAB ÄÄÄ::: ~~~^̂̂êêêÆÆƼ¼¼êêê
õ�ª
õ�ª�L«§p(x) = a1x
n + a2x
n−1 + . . .+ anx
1 + an+1 L«
� n+1 ��þ p = [a1, a2, . . . , an, an+1]¶
õ�ª\{Ú~{§$g�õ�ªXêØv�pg�^”Öv¶
õ�ª¦{ÚØ{§conv(p1,p2), [q,r]=deconv(p1,p2)
e p1 = [1, 8, 0, 0,−10], p2 = [2,−1, 3]§* conv(p1,p2),
[q,r]=deconv(p1,p2) �(J¶
õ�ª¦�§polyder(p), polyder(p1,p2), [p,q]=polyder(p1,p2)
e p1 = [1,−1], p2 = [1,−1, 3]§* p=polyder(p1),
p=polyder(p1,p2), [p,q]=polyder(p1,p2) �(J¶
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 17 / 139
MATLAB ÄÄÄ::: ~~~^̂̂êêêÆÆƼ¼¼êêê
õ�ª�¦��¦�§polyval(p,x), roots, fzero
e p=[1,-6,11,-6], x=1:4§* polyval(p,x), roots(p),
fzero(’x-10̂ x+2’,0.5) �(J”
Example 7
-Ù�±��±ê�O�úª�
Mp =
rL
12
[
1− 1
(1+r/12)12y
]
Ù¥§L´�±ê�§ü ´��¶Mp´z��±ê§ü ´�
�¶r´c|Ƕy´�±cê§ü ´c”
®���20co�17����±§��±ê�0.125��§¦c|Ç”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 18 / 139
MATLAB ÄÄÄ::: ~~~^̂̂êêêÆÆƼ¼¼êêê
¼ê4�
��¼ê4�§fminbnd(’file’,x1,x2)§*
fminbnd(’x 3̂− 2 ∗ x − 5’,0,5) �(J¶
õ�¼ê4�§fminsearch(’file’,x0)
¦õ�¼ê
f (x , y , z) = x +
y2
4x
+
z2
y
+
2
z
3 (0.5,0.5,0.5) NC����”
ïá¼ê©� f.m
function w=f(p)
x=p(1);
y=p(2);
z=p(3);
w=x+y^2/(4*x)+z^2/y+2/z;
3·-I�¥N^ fminsearch ¼ê
w=fminsearch(’f’,[0.5,0.5,0.5])
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 19 / 139
MATLAB ÄÄÄ::: ÎÎÎÒÒÒ$$$���
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 20 / 139
MATLAB ÄÄÄ::: ÎÎÎÒÒÒ$$$���
ÎÒ$�
ïáÎÒé�
ïáÎÒCþÚÎÒ~þ§sym, syms
syms a b;
a=sym(’3’);
b=sym(’4’);
x=3;
y=4;
a*a+b*b
x*x+y*y
cos((a+b)^2)
cos((x+y)^2)
ïáÎÒL�ª§|^ ’ ’, sym, syms¶
ÎÒL�ª¥Cþ�(½§findsym§ÎÒ~þØw«3(J¥”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 21 / 139
MATLAB ÄÄÄ::: ÎÎÎÒÒÒ$$$���
ÎÒL�ª$�
\!~!¦!ØÚ�$�¶
Example 8
-f = 2×2 + 3x − 5, g = x2 − x + 7§* (J”
Ϫ©)�Ðm§factor, expand¶
Example 9
-s = x3 − 6×2 + 11x − 6§* (J”
ÎÒL�ª�ê�L�ª�m�=�§eval¶
Example 10
-x =′ (1 + sqrt(5))/2′§* (J”
{zÎÒL�ª§simple, simplify, pretty”
Example 11
-s = (x3 − 4×2 + 16x)/(x3 + 64)§* (J”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 22 / 139
MATLAB ÄÄÄ::: ÎÎÎÒÒÒ$$$���
ÎÒ�È©
ÎÒ4�§limit(f,x,a), limit(f,x,a,’left’), limit(f,x,a,’right’)¶
Example 12
lim
x→0
(1 + x)
3t
x , lim
x→0−
(
e
1
x − 1
)
, lim
x→0+
(
e
1
x − 1
)
, lim
x→0
|x |
x
ÎÒ�ê§diff(s,x), diff(s,x,n)¶
Example 13
y = e−ax
2
+ x , y ′; y = cos (x2), y ′′
ÎÒÈ©§int(s,x), int(s,x,a,b)§þe�L«�{(¹¶
Example 14 ∫ 1
0
√
1− x2dx ,
∫ sin x
0
xtdt,
∫ ∞
0
e−x
2
dx
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 23 / 139
MATLAB ÄÄÄ::: ÎÎÎÒÒÒ$$$���
ÎÒ?ê§symsum(s,x,n,m)¶
Example 15
∞∑
n=1
1
n2
,
∞∑
n=1
(−1)n−1
1
2n − 1
¼ê��VÐm§taylor(f,x,a,Name,Value)¶
Example 16
(1). ¦¼êy = tan(x)3x = 0?��VÐmª�cl�¶
(2). ¦¼êy = ln(x)3x = 1?��VÐmª�cÊ�”
ÎÒ�§§[x1,. . . ,xn]=solve(s1,. . . ,sn,’x1’,. . . ,’xn’)¶
Example 17
x + ay + a2z = a3
x + by + b2z = b3
x + cy + c2z = c3
a,b,c�®��pÉ¢ê
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 24 / 139
MATLAB ÄÄÄ::: ÎÎÎÒÒÒ$$$���
Example 18 (ÎÒ�©)
�¤�>S�^´�^¼ê’u�¤x��ê”��¤¼ê©O�
u1(x) = ln x Ú u2(x) =
x1−a
1− a
�§¦�¤�>S�^u′(x)”>S�^u′(x)��5½Â�
σ(x) = −
u′′(x)
u′(x)
x .
ÁO��^¼êu1(x)Úu2(x)�>S�^��5″
Example 19
|^ rank(A,b)=rank(A); A\b; null(A); Ï)Ú�y)ÊÚ§¦)e��
5�§| {
ax1 − bx2 + bx3 = c
bx1 + ax2 + ax3 = d
.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 25 / 139
MATLAB ÄÄÄ::: ÎÎÎÒÒÒ$$$���
Example 20 (ÎÒÈ©)
e®�p = 1�§q = 4§¦I¦¼êq(p)§¦Ù÷vI¦�d��5�
�(p) = −
p
q
dq
dp
=
2p2
p2 + 1
.
Example 21 (JacobiÝ
)
�þ¼êf = (f1, f2, . . . , fm)’u�þv = (x1, x2, . . . , xn)�JacobiÝ
µ
J =
∂f
∂v
=
∂f1
∂x1
∂f1
∂x2
. . .
∂f1
∂xn
…
…
…
∂fm
∂x1
∂fm
∂x2
. . .
∂fm
∂xn
.
ef�(ax + by by)Ú(xyz y x + z)§|^jacobian¼êO�JacobiÝ
”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 26 / 139
MATLAB ÄÄÄ::: ÎÎÎÒÒÒ$$$���
|^solve·-¦)�ê�§Ú�§|
Example 22 (�g�§�))
¦f = ax2 + bx + c (a 6= 0)�ü��”
Example 23 (�§|�))
éu�§| {
10x + 12y + 16t = 0
5x − y − 13t = 0
,
¦CþtÚy’ux�)¶
eØ�½tÚy§* (J”
Example 24 (CþO�)
®�ü��§§4×2 + y2 = 4Úx2 + 4y2 = 4a§¦ëê�§�ÎÒ)§
¿|^subs¼ê¦�a = 1��ê�)”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 27 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 28 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
MATLAB ?§
§S��(�
^S(�§êâÑ\§O�?n§êâÑѶ
ÀJ(�§if �é�n«(�
(1) {ü(�
if ^�
�é|
end
(2) ��ÀJ(�
if ^�
�é|1
else
�é|2
end
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 29 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
(3) õi@(�
if ^�1
�é|1
elseif ^�2
�é|2
…
elseif ^�m
�é|m
else
�é|n
end
Example 25
y =
sin x
x
, x 6= 0,
1, x = 0.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 30 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
switch �é
õÀJ(�
switch L�ª
case L�ª1
�é|1
case L�ª2
�é|2
…
case L�ªm
�é|m
otherwise
�é|m+1
end
Ì��é§for, while, ^�þö�)¤���.Ý
¶
break Ú continue �é§aÑ��Ì��´�cÌ�”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 31 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Polynomial Interpolation
Suppose we have a set of ordered points {x0, x1, . . . , xn} and either a
continuous function f (x) defined on [x0, xn] or a set of y−values
y0, y1, . . . , yn corresponding to the x−values.
The polynomial interpolation problem is to find a polynomial p(x) of
degree at most n that interpolates the data (x0, y0), (x1, y1), . . . , (xn, yn)
such that
y0 = p(x0)
y1 = p(x1)
…
yn = p(xn)
and we say that p interpolates f (or the data) at {x0, x1, . . . , xn}, and that
p is an interpolant.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 32 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Natural Form
Let x0 < x1 < . . . < xn and y0, y1, . . . , yn be given, and let
p(x) =
n∑
i=0
aix
i = a0 + a1x + a2x
2 + . . .+ anx
n
represent the interpolant. We represent the interpolation requirements as
a0 + a1x0 + a2x
2
0 + . . .+ anx
n
0 = y0
a0 + a1x1 + a2x
2
1 + . . .+ anx
n
1 = y1
...
a0 + a1xn + a2x
2
n + . . .+ anx
n
n = yn
⇒
1 x0 . . . x
n
0
1 x1 . . . x
n
1
...
...
. . .
...
1 xn . . . x
n
n
a = y
Nevertheless, we may need to solve a highly ill-conditioned linear system,
and handle miscellaneous problems such as roundoff error, cancellation and
Runge phenomenon.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 33 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Example 26
Suppose we wish to interpolate a polynomial to the data points
(0, 0), (π/2, 1), (π, 0), (3π/2,−1) from the sine curve.
p(x) = 0.0860x3 − 0.8106x2 + 1.6977x .
Plot p(x) and sin x in the same figure.
Polynomial interpolants of high degree tend to be oscillatory, that is, they
often must make large twists and turns to go through the nodes.
Example 27 (Runge Phenomenon)
Plot the Runge function
f (x) =
1
1 + 25x2
with a polynomial interpolant of degree 10 (11 points −1 : .2 : 1).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 34 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Figure: Runge’s phenomenon
The red curve is the Runge function.
The blue curve is a 5th-order interpolating polynomial (using six
equally spaced interpolating points).
The green curve is a 9th-order interpolating polynomial (using ten
equally spaced interpolating points).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 35 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Lagrange Form
Suppose we wish to interpolate a line to two distinct points xa, xb. Define
La(x) =
x − xb
xa − xb
, Lb(x) =
x − xa
xb − xa
.
Then we have
La(xa) = 1, La(xb) = 0, Lb(xa) = 0, Lb(xb) = 1,
and both La(x) and Lb(x) are linear functions of x .
Given ya = f (xa) and yb = f (xb), we see that
` = yaLa(x) + ybLb(x)
has the desired properties and interpolates f at (xa, ya) and (xb, yb).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 36 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Given an arbitrary list of n + 1 distinct ordered nodes x0 < x1 < . . . < xn,
define the functions
Ln,i (x) =
(x − x0)(x − x1) . . . (x − xi−1)(x − xi+1) . . . (x − xn)
(xi − x0)(xi − x1) . . . (xi − xi−1)(xi − xi+1) . . . (xi − xn)
=
n∏
k=0,k 6=i
(x − xk)
(xi − xk)
, i = 0, 1, . . . , n.
Langrage polynomials are polynomials of precise degree n with the
property that Ln,i (xj) = 0 if j 6= i , and Ln,i (xi ) = 1.
Since a sum of polynomials of degree n is a polynomial of degree at most
n, given (x0, y0), (x1, y1), . . . , (xn, yn), we have
p(x) = y0Ln,0(x) + y1Ln,1(x) + . . .+ ynLn,n(x)
=
n∑
i=0
yiLn,i (x).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 37 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Example 28 (Lagrange Polynomials)
Suppose we wish to interpolate a polynomial to the data points
(0, 0), (π/2, 1), (π, 0), (3π/2,−1) from the sine curve.
p1(x) = 0.0860x
3 − 0.8106x2 + 1.6977x , (natural form, how?),
p2(x) = 0.0860x
3 − 0.8106x2 + 1.6977x , (Lagrange form, how?).
Two forms are in agreement with each other.
What is the difference between
p(x) =
n∑
i=0
aix
i and p(x) =
n∑
i=0
yiLn,i (x) ?
The natural form is easy to evaluate but hard to find.
The Lagrange form is easy to find but hard to evaluate.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 38 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Example 29 (Newton’s method)
Suppose that f (λ) = −a0 + a1λ+ a2λ2 + . . .+ anλn, where n > 1 and
ai > 0, i = 0, 1, . . . , n. Here is an iterative technique that generates a
sequence λ0, λ1, . . . , λk , . . . of estimates that converges to the root λ > 0,
solving f (λ) = 0. Start with any λ0 > 0 close to the solution and use
f ′(λk) = a1 + 2a2λk + 3a3λ
2
k + . . .+ nanλ
n−1
k , λk+1 = λk −
f (λk)
f ′(λk)
.
Locate the root of f (λ) = −1 + λ+ λ2 in [0, 1] with λ0 = 1.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 39 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Example 30
�Åß�µ?Û���u1���ê§XJ´óê§òÙر2¶XJ´
Ûê§òÙ¦±32\1§�E$�”Á¯µk¦ù�L§Øª����
êíº
Example 31
8¿üSµ´ïá38¿ö�þ��«k��üS�{“äNÚ½�
(1). ���m§¦Ù���ü�®²üSS��Ú¶
(2). �½ü���§�Ð �©O�ü�®²üSS��å© �¶
(3). ‘�ü���¤�����§ÀJ�é�����\�Ü¿�m§
¿£Ä���e� �¶
(4). EÚ½3��,�����S��¶
(5). ò,�S��e�¤k����E��Ü¿S��”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 40 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
¼ê©�
�(�
function [a1,…,am]=fname(b1,…,bn)
5º`²Ü©
¼êN�é
function [s,c]=circle(r)
% circle O��»� r ��¡ÈÚ±�
% [s,c]=circle(r)
% r ��»§s �¡È§c �±�
s=pi*r*r;
c=2*pi*r;
3·-I�¥N^ [s,c]=circle(10) Ú help circle
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 41 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
¼ê�i@N^
Example 32
Fibonacci ê�½ÂXeµ
fn = fn−1 + fn−2
f2 = 1
f1 = 1
¦ n ≥ 2 �?¿ Fibonacci ê��”
function f=fib(n)
% ^u¦ Fibonacci ê��¼ê�
% f=fib(n)
if n>2
f=fib(n-1)+fib(n-2);
else
f=1;
end
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 42 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
Example 33
(1). |^½È©�½Â§©Oæ^Ý/{ÚF/{O�½È©
I =
∫ 1
0
x2dx
(2). Á^ü©{¦�§x3 − 4×2 + 1 = 03[0, 1]þ��Cq�”
(3). |^�?êÐmª
sin x = lim
n→∞
Sn =
∞∑
n=1
an =
∞∑
n=1
(−1)n−1
x2n−1
(2n − 1)!
= x −
x3
3!
+
x5
5!
−
x7
7!
+ . . .
O�sin x�Cq�§�¦Ø��u10−8″
sin x = lim
n→∞
Sn, |sin x − Sn| ≤ an+1
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 43 / 139
MATLAB ÄÄÄ::: MATLAB ???§§§
����
Example 34
l��¯K´��±IS�Ú��µ�¯KµXÛU
38× 8�IS�
ÚÚ�þ��l���§¦�?Û����ÑÃ{��¯KÙ¦��
�º�
��d8�§?ü���ÑØU?uÓ�^î1!p1½��
þ”
�¦µ¦^êâ(�¥�Ò(stack)¢y§�ª(Jæ^Ý
/ªÑÑ”
Example 35
òê�Å�3IS�Ú�8× 8Ú��,���¥§êUrÚ5K?1£
Ä”�¦z����?\�g§rHÚ�þ�Ü64���”?��48
§S§¦Ñê�1r´�”
�¦µ¦^êâ(�¥�Ò(stack)5¢y§òêi1, 2, . . . , 64�gW\
��8× 8��
§�ª(Jæ^Ý
/ªÑÑ”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 44 / 139
MATLAB ÄÄÄ::: ���������nnn���±±±ããã
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 45 / 139
MATLAB ÄÄÄ::: ���������nnn���±±±ããã
���n�±ã
±�ü����
Example 36
|^ã�`²
lim
x→0+
tan x − sin x
x3
=
1
2
x=1:-0.00005:0.001;
y=(tan(x)-sin(x))./(x.^3);
plot(x,y,’o’)
Example 37
Á^Äxü«4 ��â”close; axis square; fill; pause.
Á^4�Ixo�p4�r = cos (2θ)”polar(theta,rho,À�).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 46 / 139
MATLAB ÄÄÄ::: ���������nnn���±±±ããã
function myfun(n)
close;
clear all;
t=0:.005:2*pi;
x=cos(t);
y=sin(t);
for side=3:n
plot(x, y, ’*’);
axis square;
hold on;
for k=1:side
theta(k)=(2*pi/side)*(k-1);
end
fill(cos(theta), sin(theta), ’r’);
pause(1);
hold off;
end
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 47 / 139
MATLAB ÄÄÄ::: ���������nnn���±±±ããã
±�õ����
Example 38
3Ó��IS±�ü^{��Ä�
y = sin
(
t +
π
4
)
Ú y = 2 sin
(
2t +
π
8
)
t=-4*pi:0.005:4*pi;
y=[sin(t+pi/4);2*sin(2*t+pi/8)];
plot(t,y)
ã/�±§|^ hold on/off ¼ê§~Xþ~3 t ��Ð��
y1=sin(t+pi/4);
plot(t,y1);
hold on;
y2=2*sin(2*t+pi/8);
plot(t,y2,’r’);
hold off;
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 48 / 139
MATLAB ÄÄÄ::: ���������nnn���±±±ããã
����ª
Example 39
3Ó��IXS§©O^ØÓ�/ÚôÚ±�¼ê
y = 0.2e−0.5x cos(4πx) Ú y = 2e−0.5x cos(πx)
�ã�§¿IÑü���:”
x=0:2*pi/1000:2*pi;
y1=0.2*exp(-0.5*x).*cos(4*pi*x);
y2=2*exp(-0.5*x).*cos(pi*x);
k=find(abs(y1-y2)<1e-2); x3=x(k); y3=0.2*exp(-0.5*x3).*cos(4*pi*x3); plot(x,y1,x,y2,’k:’,x3,y3,’ro’); ã/I5��I��µtitle, xlabel, ylabel, zlabel, text(x,y,`²), axis" Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 49 / 139 MATLAB ÄÄÄ::: ���������nnn���±±±ããã Û¼ê�ãµezplot¶ subplot(2,2,1); ezplot(’sin(x)’); subplot(2,2,2); ezplot(’x^2+y^2-1’,[-1,1],[-1,1]); subplot(2,2,3); ezplot(’x^3+y^3-5*x*y’); subplot(2,2,4); ezplot(’2*cos(t)’,’sin(t)’,[0,2*pi]); Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 50 / 139 MATLAB ÄÄÄ::: ���������nnn���±±±ããã n�ã/ n��µplot3 Example 40 ±�Ú^� x = cos t y = sin t z = t (0 ≤ t ≤ 4π) t=0:0.001:4*pi; x=cos(t); y=sin(t); z=t; plot3(x,y,z); grid on; Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 51 / 139 MATLAB ÄÄÄ::: ���������nnn���±±±ããã n�¡µ�n�k²¡§2�m" Example 41 ±�²¡ z = x + y Ú^=�Ô¡ z = x2 + y2 �ã/" x=0:1:2; y=0:1:2; [x,y]=meshgrid(x,y); z=x+y; axis([0,2,0,2,0,4]); mesh(x,y,z); x=-8:0.1:8; y=-8:0.1:8; [x,y]=meshgrid(x,y); z=x.^2+y.^2; surf(x,y,z); Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 52 / 139 MATLAB ÄÄÄ::: ���������nnn���±±±ããã Example 42 �Ñ^S ��õ>/¡È%C�¡È�ê�¢��ã/ü«”
sn =
1
2
r2n sin
2π
n
, Sn = r
2n tan
π
n
, (n > 2).
Example 43
��y = x(1− x)3[0, 1]þ�=Ä��§lAÛþ`²Y²����
35″
Example 44
®�¼ê
z =
1
1 + D2n(x , y)/D0
, D(x , y) =
√
(x − x0)2 + (y − y0)2
D0 = 200, n = 2, x0 = y0 = 16§^meshÚsurf±�n�ã�”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 53 / 139
���`̀̀zzz¯̄̄KKK¦¦¦)))
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 54 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 55 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
�55y¯K�IO.
max f (x) = c1x1 + c2x2 + . . .+ cnxn
s.t.
a11x1 + a12x2 + . . .+ a1nxn = b1
a21x1 + c2x22 + . . .+ cnx2n = b2
. . .
am1x1 + am2x2 + . . .+ amnxn = bm
xi ≥ 0, i = 1, 2, . . . , n
½
max f (x) = cTx
s.t. Ax = b, x ≥ 0.
Example 2.1
òe��55y�.IOz
min x1 − 2×2 + x3
s.t.
x1 + x2 + x3 ≤ 3
x1 + x2 − 2×3 ≥ 1
− x1 + 2×2 + 3×3 = 4
, x1 ≥ 0, x2 ≥ 0.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 56 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
�55y¯K¥)�Vg
Ä�)µl A ¥ÀÑ�ÛÉÝ
B ¦� BxB + NxN = b¶
�1)§�1�µ÷v D1 = {x |Ax = b, x ≥ 0} �¤k)¶
Ä��1)µ D = {x |Ä�) ∩ �1)}¶
�`)µ¦8I¼ê������Ä��1)”
Example 2.2
max − x1 + 2×2 − x3
s.t.
{
x1 + x2 + x3 = 3
x1 − x2 − 2×3 = 1
, x1, x2, x3 ≥ 0
A =
(
1 1 1
1 −1 −2
)
K B1 =
(
P1 P2
)
⇒ xB1 =
(
2 1 0
)
B2 =
(
P2 P3
)
⇒ xB2 =
(
0 7 −4
)
B3 ⇒ xB3 =?
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 57 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
�55y¯K�MATLAB¦)�{
�55y¯K�MATLABIO.
min cT x
s.t. Ax ≤ b, Aeqx = beq, lb ≤ x ≤ ub
Example 2.3
�Ñe��55y�.¥� c ,A, b,Aeq, beq, lb, ub
max 4×1 − 2×2 + x3
s.t.
2×1 − x2 + x3 ≤ 12
− 8×1 + 2×2 − 2×3 ≥ 8
− 2×1 + x3 = 3
x1 + x2 = 7
x1, x2, x3 ≥ 0
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 58 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
�55y¯K¦)�MATLAB¼êN^
x = linprog(c ,A, b)
x = linprog(c ,A, b,Aeq, beq)
x = linprog(c ,A, b,Aeq, beq, lb, ub)
x = linprog(c ,A, b,Aeq, beq, lb, ub, x0)
x = linprog(c ,A, b,Aeq, beq, lb, ub, x0, options)
[x , fval ] = linprog(. . .)
[x , fval , exitflag ] = linprog(. . .)
[x , fval , exitflag , output] = linprog(. . .)
[x , fval , exitflag , output, lambda] = linprog(. . .)
Example 2.4
Á^þã·-¦)þ~¥��55y�.”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 59 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
�55y¢~
Example 2.5
,ó�I�)�AÚBü«�¬±÷v½|I¦”ùü«�¬�)�þ
I�²Lü�ó²6§”z)�1ú6�A�¬31��Ú1��ó²�
6§Ñ�©O�4��Ú6��¶z)�1ú6�B�¬331��Ú1�
�ó²�6§Ñ�©O�6��Ú8��”du)�Oy��¦§�ø1
��Ú1��ó²�6§ó�©O�240��Ú480��”Tó�3)
�B�¬�Ó�§¬�ÑB�¬C§z)�1ú6�B�¬¬�)2ú6�
B�¬C§
ØI \?Û¤^§duB�¬C�|^ǯK§¦
�C¥�Ü©�J|§Ù§Ü©�U�¢”
�âØ�§ÑÈ1ú6A�¬�±J|600�§ÑÈ1ú6B�¬�±J
|1000�§ÑÈ1ú6C�¬�±J|300�§
�¢1ú6C�¬I�º
�200 �”²½|ýÿ§3OyÏS§�¬C����Èþ�50 ú6″
KA�XÛSüAÚBü«�¬��þ§âU¦ó��oJ|���
pº
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 60 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
Example 2.6
,Å�yPk]�200��§�
¼����ÂçTÅ�û½ò
ù200��?1Ý]§±Ï��£�”yko«�Y�øÀJ§Ý]�
ª�zccÐòÅ�±k�¤k]�Ñ^uÝ]”
1 l11c�14c�zccÐÑI�Ý]§gcc”£�|1.15¶
2 13ccÐÝ]§15cc”£�|1.25§��Ý]��80��¶
3 12ccÐÝ]§15cc”£�|1.40§��Ý]��60��¶
4 zccÐÝ]§zcc”£�|1.06″
KAæ^Û«Ý]|Üüѧ¦�TÅ�15cc”o]���º
11c 12c 13c 14c 15c
�Y� x11 x21 x31 x41
�Y� x32
�Yn x23
�Yo x14 x24 x34 x44 x54
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 61 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
Example 2.7
�kü�ïá�AÚB§zcâ���þ©O�35�ëÚ55�ë§ù
â��øA�`!¯!Zn�ïÓó/§z�ïÓó/éâ��I¦þ
©O�26�ë!38�ëÚ26�ë§�ïá��ïÓó/�m�$¤£�
�/�ë¤XeL¤«§KA�XÛN$âU¦�o$¤��º
ó/` ó/¯ ó/Z
ïá�A 10 12 9
ïá�B 8 11 13
ó/` ó/¯ ó/Z ÑÑoþ
ïá�A x11 x12 x13 35
ïá�B x21 x22 x21 55
�Âoþ 26 38 26 90
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 62 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
Example 2.8
,r�|k5«��”®��«r��ü d�Úzü r���x
�!¶Ô�!�)�¹þXeL¤«§q�T|zF��I��x
�70ü !¶Ô�3ü !�)�10Îü “¯XÛ·ÜN�ù5«r
�§âU¦o¤��$º
r�«a �x�/ü ¶Ô�/ü �)�/Îü üd
1 0.30 0.10 0.05 2
2 2.20 0.05 0.10 7
3 1.00 0.02 0.02 4
4 0.60 0.20 0.20 3
5 1.80 0.05 0.08 5
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 63 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���555555yyy
Example 2.9
¦)Xe`z¯K
min |x |+ |y |+ |z |,
s.t.
{
x + y ≤ 1,
2x + z = 3.
|^eãC�
m =
x + |x |
2
, n =
|x | − x
2
,
ò��55y¯K=z¤�55y¯K¦)”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 64 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 65 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
�ê5y�êÆ�.
max cT x ,
s.t.
{
Ax = b,
x ≥ 0, ��ܽöÜ©��ê�.
�y©�X�ê5y§·Ü�ê5yÚ0–15y”
�ê5y¯K� MATLAB ¦)�{
�5·Ü�ê5y¯K
[x , fval , exitflag ] = intlinprog(c , intcon,A, b,Aeq, beq, lb, ub)
min cT x ,
s.t.
Ax ≤ b,
Aeqx = beq,
lb ≤ x ≤ ub,
xi ≥ 0, i = 1, 2, . . . , n,
xj��ê�, j ∈ intcon.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 66 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
[x , fval , exitflag ] = intprog(c ,A, b,Aeq, beq, lb, ub,M,TolXInteger)
min cT x ,
s.t.
Ax ≤ b,
Aeqx = beq,
lb ≤ x ≤ ub,
xi ≥ 0, i = 1, 2, . . . , n,
xj��ê�, j ∈ M.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 67 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
Example 2.10
max x1 + x2,
s.t.
4×1 − 2×2 ≥ 1,
4×1 + 2×2 ≤ 11,
2×2 ≥ 1,
x1, x2 ≥ 0���ê�.
c=[-1;-1];
A=[-4 2;4 2;0 -2];
b=[-1;11;-1];
lb=[0;0];
intcon=[1;2];
[x,fval]=linprog(c,A,b,[],[],lb,[])
[x1,fval1]=intlinprog(c,intcon,A,b,[],[],lb,[])
[x2,fval2]=intprog(c,A,b,[],[],lb,[],M,Tol)
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 68 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
^ MATLAB ¦) 0–1 �5�ê5y¯K
min cT x ,
s.t.
Ax ≤ b,
Aeqx = beq,
x = 0 or 1.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 69 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
Example 2.11
éu 0–1 5y�.
max 20×1 + 6×2 + 8×3 + 9×4,
s.t.
10×1 + 6×2 + 5×3 + 2×4 ≤ 19,
7×1 + 2×2 + 2×3 + 4×4 ≤ 11,
2×1 + x2 + x3 + 10×4 ≤ 12,
x4 ≤ x2 + x3,
x1, x2, x3, x4 = 0 or 1.
�©O^ linprog Ú intlinprog (½ intprog )¦)”
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 70 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
Example 2.12
��À1ö�
Ñ1I��nÙ1o§�´¨����õ�U«É25ú
6�Ô¬”TÀ1ö�k6�Ô¬�ø��§Ô¬�éA?Ò§�þ4
Ùd�XeL¤«”du�����§TÀ1öû½��¦�Upd�
�Ô¬§@o¦ATXÛÀJ�¦��Ô¬�od���º
Ô¬?Ò 1 2 3 4 5 6
�þ/ú6 4.8 6.2 5.7 3.6 4.4 8.5
d�/� 120 180 150 100 90 230
Example 2.13
,�û8c�k�«5��A�g+§Ù�½�Ý�17 �”y3I�
�â�ré�Ý�I¦éÙ)��g+?1��§XJ�rI
�20�8��!40�6��!80�4���T«A�g+§KA�XÛé
g+��âU¦^���£I²(XÛ½Â^���¤º
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 71 / 139
Xi CHEN
obj=660
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
Example 2.14
,ó�)�,«�¬±÷v½|I¦”²L½|NïÚÚO§ýOT«
>ì3�5o�GÝ�½|I¦þ©O�1500�§2000�§4000�
Ú1000�”XJT�û½3,�GÝmó§I�ó§O�¤20000 �§
z�Åì�)�¤��500�§XJ÷v½|I¦±�GÝ”k¥��
¬�{§Kz�>ì��;¤�10�”b½ó�ÃЩ¥�§A�XÛ
Sü)�§âU3÷v½|I¦�^�e¦�o¤^��?
Example 2.15
,ë£U,3�U���mã¤IÑÖ
�<êXeL¤«"b�ÑÖ
þ���m2��mã�m©��§¿�ëYó�8��"KTU,
A����õ�¶ÑÖ
º
�g 1 2 3 4 5 6
�mã 08–12 12–16 16–20 20–24 00–04 04–08
I¦<ê 100 120 80 60 30 50
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 72 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
©{½.{Ú½£��z8I¼ê¤
(1). Щ¦)�ê5y�tµ¯Kµ¦)Ùtµ�55y§e��ê
)§=��ê5y��`)"ÄK§Ð©e.��−∞"
(2). ïá©{äµ3?Û��£f¤¯K¥§lØ÷v�ê�¦�Cþ
¥ÀÑ��?1?n§ÏL\\�ép½��åò��£f¤¯K
©|�ü�É�?�Ú�å�f¯K§ �|¢�«�"dd§f
¯KeØ÷v�ê�¦K?�Ú�e?1©{§/¤��©{ä"
(3). ½.�}{µÏLØä/©{Ú¦)��f¯K§©{½.{òØ
ä?�d®²����`�ê)(½�e."Ù¥§¦)f¯K�
UÑy±e(Jµ
Ã�1)§ÃLUY©{¶
�����ê)§ÃLUY©{§�#e.¶
������ê)§À8I¼ê���¹û½}{�´UY©{"
(4). UìþãÚ½S�§z�e.�?U±�§A�u�¤k�vk¦
)L�f¯K¿}�@
8I¼ê��u#�e.�f¯K"
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 73 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���êêê555yyy
Example 2.16
^©|½.{¦)�ê5y¯K
max x1 + x2,
s.t.
4x1 − 2x2 ≥ 1,
4x1 + 2x2 ≤ 11,
2x2 ≥ 1,
x1, x2 ≥ 0���ê�.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 74 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
1 MATLAB Ä:
Ý
�ê|$�
~^êƼê
ÎÒ$�
MATLAB ?§
���n�±ã
2 �`z¯K¦)
�55y
�ê5y
�g5y
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 75 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
�g5y¯K�êÆ�.
min
1
2
xtHx + ctx,
s.t. Ax ≥ b.
Ù¥§H ∈ Rn×n�n�¢é¡Ý
"
¦¦¦)))���{{{
����{£�k�ª�夶
.�KF¦f{£�k�ª�夶
f�m&6�{£�k�ª�å½ö�kþe.�夶
k�8�{¶
Wolfe�{¶
Lemke�{"
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 76 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
����{µòA©)¤A = (B,N)§Ù¥B�ÄÝ
§�A/
òx, c,H�Xe©¬µ
x =
(
xB
xN
)
, c =
(
cB
cN
)
, H =
(
H11 H12
H21 H22
)
,
K��BxB + NxN = b§=xB = B
−1b− B−1NxN§�\�g5y��
min φ(x) =
1
2
xtNĤ2xN + ĉ
t
NxN .
Ù¥
Ĥ2 = H22 −H21B−1N−Nt
(
B−1
)t
H12 + N
t
(
B−1
)t
H11B
−1N,
ĉN = cN −Nt
(
B−1
)t
cB +
[
H21 −Nt
(
B−1
)t
H11
]
B−1b.
eĤ2�½§��Ã�å¯K�`)�
x∗N = −Ĥ
−1
2 ĉN ⇒ x
∗ =
(
x∗B
x∗N
)
=
(
B−1b
0
)
+
(
B−1N
−I
)
Ĥ−12 ĉN .
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 77 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
`:µg�{ü²
§¦^�B¶
":µB�U�C��ÛÉ�
§l
Úåx∗�ê�ؽ"
Example 2.17
^����{¦)�g5y¯Kµ
min x21 + x
2
2 + x
2
3 ,
s.t.
{
x1 + 2x2 − x3 = 4,
x1 − x2 + x3 = −2.
x∗ = (x∗1 , x
∗
2 , x
∗
3 )
t =
(
2
7
,
10
7
,−
6
7
)t
.
�dAtλ∗ = ∇f (x∗) = Hx∗ + c ��
λ∗ = (λ∗1, λ
∗
2)
t =
(
8
7
,−
4
7
)t
.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 78 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Example 2.18
.�KF¦f{�nµ
L(x,λ) =
1
2
xtHx + ctx− λt(Ax− b),{
∇xL(x,λ) = 0,
∇λL(x,λ) = 0,
⇒
(
H −At
−A 0
)(
x
λ
)
= −
(
c
b
)
.
=K-T^�§Ù¥�.�KFÝ
é¡�Ø�½�½¶e�½§Kk(
H −At
−A 0
)−1
=
(
Q −R
−Rt G
)
,
Q = H−1 −H−1At(AH−1At)−1AH−1 = H−1 − RAH−1,
R = H−1At(AH−1At)−1 = −H−1AtG,
G = −(AH−1At)−1.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 79 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
�x(0)´?��1)§=÷vAx(0) = b§K3x(0)?�8I¼êFÝ�±
L«�∇f
(
x(0)
)
= Hx(0) + c"Kd{
x∗ = −Qc + Rb,
λ∗ = Rtc− Gb,
⇒
{
x∗ = x(0) −Q∇f
(
x(0)
)
,
λ∗ = Rt∇f
(
x(0)
)
.
Example 2.19
^.�KF{¦)�g5y¯Kµ
min x21 + 2x
2
2 + x
2
3 − 2x1x2 + x3,
s.t.
{
x1 + x2 + x3 = 4,
2x1 − x2 + x3 = 2.
x∗ =
(
21
11
,
43
22
,
3
22
)t
.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 80 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
�g5y¯K�MATLAB¦)�{
�g5y¯K�MATLABIO.
min 1
2
xTHx + cT x
s.t. Ax ≤ b, Aeqx = beq, lb ≤ x ≤ ub
�55y¯K¦)�MATLAB¼êN^
x = quadprog(H, c ,A, b)
x = quadprog(H, c ,A, b,Aeq, beq)
x = quadprog(H, c ,A, b,Aeq, beq, lb, ub)
x = quadprog(H, c ,A, b,Aeq, beq, lb, ub, x0)
[x , fval ] = quadprog(. . .)
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 81 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Example 2.20
Á^þã·-¦)þ~¥��g5y¯K"
Example 2.21
¦)±e�g5y¯Kµ
min
1
2
x21 + x
2
2 − x1x2 − 2x1 − 6x2,
s.t.
x1 + x2 ≤ 2,
− x1 + 2x2 ≤ 2,
2x1 + x2 ≤ 3,
x1, x2 ≥ 0.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 82 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Definition 2.1
A set Ω in Rn is said to be convex if for every x1, x2 ∈ Ω and every real
number α, 0 < α < 1, the point αx1 + (1− α)x2 ∈ Ω.
Definition 2.2
A function f defined on a convex set Ω is said to be convex if, for every
x1, x2 ∈ Ω and every α, 0 ≤ α ≤ 1, there holds
f (αx1 + (1− α)x2) ≤ αf (x1) + (1− α)f (x2).
If, for every α, 0 < α < 1, and x1 6= x2, there holds
f (αx1 + (1− α)x2) < αf (x1) + (1− α)f (x2).
Then f is said to be strictly convex.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 83 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Definition 2.3
A function g defined on a convex set Ω is said to be concave if the
function −g is convex, and be strictly concave if −g is strictly convex.
Definition 2.4
If f ∈ C 1 is a real-valued function on Rn, f (x) = f (x1, x2, . . . , xn), define
the gradient of f to be the vector
∇f (x) =
[
∂f (x)
∂x1
,
∂f (x)
∂x2
, . . . ,
∂f (x)
∂xn
]
.
If f ∈ C 2, then define the Hessian of f at x to be the n × n matrix
F(x) =
[
∂2f (x)
∂xi∂xj
]
.
∂2f (x)
∂xi∂xj
=
∂2f (x)
∂xj∂xi
⇒ F(x) = [F(x)]t .
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 84 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.1
Let f1 and f2 be convex functions on the convex set Ω. Then the functions
f1 + f2 and af (a > 0) are both convex on Ω.
Proof.
Let x1, x2 ∈ Ω, and 0 < α < 1. Then f1(αx1 + (1− α)x2) + f2(αx1 + (1− α)x2) ≤α[f1(x1) + f2(x1)] + (1− α)[f1(x2) + f2(x2)]. The conclusion about the function af (a > 0) is immediate.
Proposition 2.2
Let f1, f2, . . . , fn be convex functions over the convex set Ω. Then a
positive combination of fi , namely,
∑n
i=1 ai fi = a1f1 + a2f2 + . . .+ anfn
(ai > 0), is again convex.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 85 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.1
Let f1 and f2 be convex functions on the convex set Ω. Then the functions
f1 + f2 and af (a > 0) are both convex on Ω.
Proof.
Let x1, x2 ∈ Ω, and 0 < α < 1. Then f1(αx1 + (1− α)x2) + f2(αx1 + (1− α)x2) ≤α[f1(x1) + f2(x1)] + (1− α)[f1(x2) + f2(x2)]. The conclusion about the function af (a > 0) is immediate.
Proposition 2.2
Let f1, f2, . . . , fn be convex functions over the convex set Ω. Then a
positive combination of fi , namely,
∑n
i=1 ai fi = a1f1 + a2f2 + . . .+ anfn
(ai > 0), is again convex.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 85 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.3
Let f ∈ C 1. Then f is convex over a convex set Ω if and only if
f (y) ≥ f (x) +∇f (x)(y − x) for all x, y ∈ Ω.
Proof.
“⇐” Since f is convex, then for all α, 0 ≤ α ≤ 1,
f (αy + (1− α)x) ≤ αf (y) + (1− α)f (x),
which indicates that for all α, 0 < α ≤ 1, f (x + α(y − x))− f (x) α ≤ f (y)− f (x), lim α→0 [f (x + α(y − x))− f (x)](y − x) α(y − x) = ∇f (x)(y − x) ≤ f (y)− f (x). Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 86 / 139 ���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy Proposition 2.3 Let f ∈ C 1. Then f is convex over a convex set Ω if and only if f (y) ≥ f (x) +∇f (x)(y − x) for all x, y ∈ Ω. Proof. “⇐” Since f is convex, then for all α, 0 ≤ α ≤ 1, f (αy + (1− α)x) ≤ αf (y) + (1− α)f (x), which indicates that for all α, 0 < α ≤ 1, f (x + α(y − x))− f (x) α ≤ f (y)− f (x), lim α→0 [f (x + α(y − x))− f (x)](y − x) α(y − x) = ∇f (x)(y − x) ≤ f (y)− f (x). Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 86 / 139 ���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy continued. “⇒” Assume f (y) ≥ f (x) +∇f (x)(y − x) for all x, y ∈ Ω. Fix x1, x2 ∈ Ω and α (0 ≤ α ≤ 1). Setting x = αx1 + (1− α)x2, we have f (x1) ≥ f (x) +∇f (x)(x1 − x), (1) f (x2) ≥ f (x) +∇f (x)(x2 − x). (2) Multiplying (1) by α and (2) by 1− α and adding, we obtain αf (x1) + (1− α)f (x2) ≥ f (x) +∇f (x)(αx1 + (1− α)x2 − x). Substituting x = αx1 + (1− α)x2, we have αf (x1) + (1− α)f (x2) ≥ f (αx1 + (1− α)x2). Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 87 / 139 ���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy The original definition essentially states that linear interpolation between two points overestimates the function. The above proposition states that linear approximation based on the local derivative underestimates the function. Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 88 / 139 ���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy Theorem 45 A quadratic form f (x) = xtAx is positive definite if and only if leading principal minors of A are all positive. It is negative definite if and only if{ |Ak | < 0, if k is odd, |Ak | > 0, if k is even.
Proof.
We skip the proof for positive definiteness. For negative definiteness, since
f (x) = f (x1, x2, . . . , xn) = x
tAx
is negative definite, we have
−f (x) = −f (x1, x2, . . . , xn) = −xtAx = xt(−A)x
is positive definite. The conclusion is immediate.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 89 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 45
A quadratic form f (x) = xtAx is positive definite if and only if leading
principal minors of A are all positive. It is negative definite if and only if{
|Ak | < 0, if k is odd, |Ak | > 0, if k is even.
Proof.
We skip the proof for positive definiteness. For negative definiteness, since
f (x) = f (x1, x2, . . . , xn) = x
tAx
is negative definite, we have
−f (x) = −f (x1, x2, . . . , xn) = −xtAx = xt(−A)x
is positive definite. The conclusion is immediate.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 89 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.1 (Taylor’s Theorem)
If f ∈ C 1 in a region containing the line segment [x1, x2], then there is an
α (0 ≤ α ≤ 1) such that
f (x2) = f (x1) +∇f (αx1 + (1− α)x2)(x2 − x1).
Furthermore, if f ∈ C 2, then there is a α (0 ≤ α ≤ 1) such that
f (x2) = f (x1) +∇f (x1)(x2 − x1)
+
1
2
(x2 − x1)tF(αx1 + (1− α)x2)(x2 − x1),
where F denotes the Hessian of f .
Definition 2.5
Given x ∈ Ω, a vector d is a feasible direction at x if there is an α such
that x + αd ∈ Ω for all α (0 ≤ α ≤ α).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 90 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.2
Let f ∈ C 2 be quadratic function. Then f is convex over a convex set Ω
containing an interior point if and only if the Hessian matrix F of f is
positive semidefinite throughout Ω.
Proof.
By Taylor’s theorem, for some α (0 ≤ α ≤ 1),
f (y) = f (x) +∇f (x)(y − x) +
1
2
(y − x)tF(x + α(y − x))(y − x). (3)
Since f is a quadratic function, (3) reduces to
f (y) = f (x) +∇f (x)(y − x) +
1
2
(y − x)tF(y − x).
By Proposition 2.3, the conclusion is immediate.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 91 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.2
Let f ∈ C 2 be quadratic function. Then f is convex over a convex set Ω
containing an interior point if and only if the Hessian matrix F of f is
positive semidefinite throughout Ω.
Proof.
By Taylor’s theorem, for some α (0 ≤ α ≤ 1),
f (y) = f (x) +∇f (x)(y − x) +
1
2
(y − x)tF(x + α(y − x))(y − x). (3)
Since f is a quadratic function, (3) reduces to
f (y) = f (x) +∇f (x)(y − x) +
1
2
(y − x)tF(y − x).
By Proposition 2.3, the conclusion is immediate.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 91 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.3
Let f ∈ C 2. Then f is convex over a convex set Ω containing an interior
point if and only if the Hessian matrix F of f is positive semidefinite
throughout Ω.
Proof.
By Taylor’s theorem, for some α (0 ≤ α ≤ 1),
f (y) = f (x) +∇f (x)(y − x) +
1
2
(y − x)tF(x + α(y − x))(y − x). (4)
“⇒” Clearly, if the Hessian is everywhere positive semidefinite, we have
f (y) ≥ f (x) +∇f (x)(y − x), (5)
which in view of Proposition 2.3 implies that f is convex.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 92 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.3
Let f ∈ C 2. Then f is convex over a convex set Ω containing an interior
point if and only if the Hessian matrix F of f is positive semidefinite
throughout Ω.
Proof.
By Taylor’s theorem, for some α (0 ≤ α ≤ 1),
f (y) = f (x) +∇f (x)(y − x) +
1
2
(y − x)tF(x + α(y − x))(y − x). (4)
“⇒” Clearly, if the Hessian is everywhere positive semidefinite, we have
f (y) ≥ f (x) +∇f (x)(y − x), (5)
which in view of Proposition 2.3 implies that f is convex.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 92 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
continued.
“⇐” Now suppose the Hessian is not positive semidefinite at some point
x ∈ Ω.
By the continuity of the Hessian it can be assumed, without loss of
generality, that x is an interior point of Ω. There is a y ∈ Ω such that
(y − x)tF(x)(y − x) < 0.
Again by the continuity of the Hessian, y may be selected so that for all
α (0 ≤ α ≤ 1),
(y − x)tF(x + α(y − x))(y − x) < 0.
By (4), (5) does not hold; which in view of Proposition 2.3 implies that f
is not convex.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 93 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.4
Let f be a convex function defined on the convex set Ω. Then the set Γ
where f achieves its minimum is convex, and any relative minimum of f is
a global minimum.
Proof.
If f has no relative minima the theorem is valid by default. Otherwise
Γ = {x : x ∈ Ω, f (x) ≤ c}. Let x1, x2 ∈ Γ. Then for α (0 < α < 1),
f (αx1 + (1− α)x2) ≤ αf (x1) + (1− α)f (x2) ≤ c .
Suppose now that x∗ ∈ Ω is a relative minimum point of f , but that there
is another point y ∈ Ω with f (y) < f (x∗). On the line αy + (1− α)x∗,
f (αy + (1− α)x∗) ≤ αf (y) + (1− α)f (x∗) < f (x∗),
contradicting the fact that x∗ is a relative minimum point.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 94 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.4
Let f be a convex function defined on the convex set Ω. Then the set Γ
where f achieves its minimum is convex, and any relative minimum of f is
a global minimum.
Proof.
If f has no relative minima the theorem is valid by default. Otherwise
Γ = {x : x ∈ Ω, f (x) ≤ c}. Let x1, x2 ∈ Γ. Then for α (0 < α < 1),
f (αx1 + (1− α)x2) ≤ αf (x1) + (1− α)f (x2) ≤ c .
Suppose now that x∗ ∈ Ω is a relative minimum point of f , but that there
is another point y ∈ Ω with f (y) < f (x∗). On the line αy + (1− α)x∗,
f (αy + (1− α)x∗) ≤ αf (y) + (1− α)f (x∗) < f (x∗),
contradicting the fact that x∗ is a relative minimum point.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 94 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.4 (First-order necessary conditions)
Let Ω be a subset of Rn and let f ∈ C 1 be a function on Ω. If x∗ is a
relative minimum point of f over Ω, then for any feasible direction d ∈ Rn
at x∗, we have ∇f (x∗)d ≥ 0.
Proof.
For any α (0 ≤ α ≤ α), the point x(α) = x∗ + αd ∈ Ω. Define the
function g(α) = f (x(α)). Then g(α) has a relative minimum at α = 0.
By the ordinary calculus we have
g(α)− g(0) = g ′(0)α + o(α),
where o(α) denotes terms that go to zero faster than α. If g ′(0) < 0,
then, for sufficiently small values of α > 0, the right side will be negative,
and hence g(α)− g(0) < 0, which contradicts the minimal nature of g(0).
Thus g ′(0) = ∇f (x∗)d ≥ 0.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 95 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.4 (First-order necessary conditions)
Let Ω be a subset of Rn and let f ∈ C 1 be a function on Ω. If x∗ is a
relative minimum point of f over Ω, then for any feasible direction d ∈ Rn
at x∗, we have ∇f (x∗)d ≥ 0.
Proof.
For any α (0 ≤ α ≤ α), the point x(α) = x∗ + αd ∈ Ω. Define the
function g(α) = f (x(α)). Then g(α) has a relative minimum at α = 0.
By the ordinary calculus we have
g(α)− g(0) = g ′(0)α + o(α),
where o(α) denotes terms that go to zero faster than α. If g ′(0) < 0,
then, for sufficiently small values of α > 0, the right side will be negative,
and hence g(α)− g(0) < 0, which contradicts the minimal nature of g(0).
Thus g ′(0) = ∇f (x∗)d ≥ 0.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 95 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Corollary 2.1
Let Ω be a subset of Rn, and let f ∈ C 1 be a function on Ω. If x∗ is a
relative minimum point of f over Ω and if x∗ is an interior point of Ω,
then ∇f (x∗) = 0.
Solve n equations in n unknowns system to derive the solution?
Example 2.22
Consider the problem
min x21 − x1 + x2 + x1x2,
s.t. x1, x2 ≥ 0.
The problem has a global minimum at x∗ = (1/2, 0)t . At x∗, the partial
derivatives, ∇f (x∗) = (0, 3/2), do not both vanish, but ∇f (x∗)d ≥ 0 for
all feasible direction d ∈ R2 at x∗ = (1/2, 0)t .
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 96 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.5 (Second-order necessary conditions)
Let Ω be a subset of Rn and let f ∈ C 2 be a function on Ω. If x∗ is a
relative minimum point of f over Ω, then for any feasible direction d ∈ Rn
at x∗, we have
i). ∇f (x∗)d ≥ 0;
ii). if ∇f (x∗)d = 0, then dtF(x∗)d ≥ 0.
Proof.
The first condition is just Proposition 2.4, and the second condition
applies only if ∇f (x∗)d = 0. Introduce x(α) = x∗ + αd and define
g(α) = f (x(α)) as before. Then in view of g ′(0) = 0, we have
g(α)− g(0) =
1
2
g ′′(0)α2 + o(α2). (6)
If g ′′(0) < 0, the right side of (6) is negative for sufficiently small α, which
contradicts the relative minimum nature of g(0).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 97 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.5 (Second-order necessary conditions)
Let Ω be a subset of Rn and let f ∈ C 2 be a function on Ω. If x∗ is a
relative minimum point of f over Ω, then for any feasible direction d ∈ Rn
at x∗, we have
i). ∇f (x∗)d ≥ 0;
ii). if ∇f (x∗)d = 0, then dtF(x∗)d ≥ 0.
Proof.
The first condition is just Proposition 2.4, and the second condition
applies only if ∇f (x∗)d = 0. Introduce x(α) = x∗ + αd and define
g(α) = f (x(α)) as before. Then in view of g ′(0) = 0, we have
g(α)− g(0) =
1
2
g ′′(0)α2 + o(α2). (6)
If g ′′(0) < 0, the right side of (6) is negative for sufficiently small α, which
contradicts the relative minimum nature of g(0).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 97 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Corollary 2.2
Let x∗ be an interior point of the set Ω, and suppose x∗ is a relative
minimum point over Ω of the function f ∈ C 2. Then
i). ∇f (x∗) = 0;
ii). For all d, dtF(x∗)d ≥ 0.
Example 2.23
Consider the problem
min f (x1, x2) = x
2
1 − x1x2 + x
2
2 − 3x2.
There are no constraints, so Ω = R2. Setting the partial derivatives of f
equal to zero yields the two equations, which have the solution x = (1, 2)t ,
the global minimum point of f .
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 98 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Example 2.24
Consider the problem
min x31 − x
2
1x2 + 2x
2
2 ,
s.t. x1, x2 ≥ 0.
Assume the solution is in the interior of Ω. By Corollary 2.2 i), we have
x∗ = (6, 9)t . Nevertheless, x∗ is not a relative minimum point since
according to
F(x) =
(
6x1 − 2x2 −2x1
−2x1 4
)
,
we have
F(x∗) =
(
18 −12
−12 4
)
,
whose determinant is −72, and thus F(x) is not positive semidefinite at x∗.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 99 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Example 2.25 (Approximation)
Suppose that through an experiment the value of a function g is observed
at m points, x1, x2, . . . , xm. Thus, g(x1), g(x2), . . . , g(xm) are known. We
wish to approximate the function by a polynomial
h(x) =
n∑
i=0
aix
i = anx
n + an−1x
n−1 + . . .+ a1x + a0
of degree n (or less), where n < m. Corresponding to any choice of the
approximating polynomial, there will be a set of errors �k = g(xk)− h(xk).
We define the best approximation as the polynomial that minimizes the
sum of the squares of these errors; i.e., minimizes
f (a) =
m∑
k=1
(�k)
2 =
m∑
k=1
(g(xk)− [an(xk)n + . . .+ a1xk + a0])
2
.
with respect to a = (a0, a1, . . . , an) to find the best coefficients.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 100 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Example 2.26 (continued)
To find a compact representation for this objective, define
qij =
m∑
k=1
(xk)
i+j , bj =
m∑
k=1
g(xk)(xk)
j , c =
m∑
k=1
g(xk)
2.
Then after a bit of algebra it can be shown that
f (a) = atQa− 2bta + c ,
where Q = (qij) and b = (b1, b2, . . . , bn+1)
t .
The first-order necessary conditions state that the gradient of f must
vanish. This leads directly to the system of n + 1 equations
Qa = b,
which can be solved to determine a.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 101 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.6 (Second-order sufficient conditions)
Let f ∈ C 2 be a function defined on a region in which the point x∗ is an
interior point. Suppose in addition that
i). ∇f (x∗) = 0;
ii). F(x∗) is positive definite.
Then x∗ is a strict relative minimum point of f .
Proof.
Since F(x∗) is positive definite, there exists an a > 0 such that for all d,
dtF(x∗)d ≥ a‖d‖2. Thus, by Taylor’s Theorem, we have
f (x∗ + d)− f (x∗) =
1
2
dtF(x∗)d + o(‖d‖2) ≥
a
2
‖d‖2 + o(‖d‖2),
which implies that both sides are positive for small d.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 102 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.6 (Second-order sufficient conditions)
Let f ∈ C 2 be a function defined on a region in which the point x∗ is an
interior point. Suppose in addition that
i). ∇f (x∗) = 0;
ii). F(x∗) is positive definite.
Then x∗ is a strict relative minimum point of f .
Proof.
Since F(x∗) is positive definite, there exists an a > 0 such that for all d,
dtF(x∗)d ≥ a‖d‖2. Thus, by Taylor’s Theorem, we have
f (x∗ + d)− f (x∗) =
1
2
dtF(x∗)d + o(‖d‖2) ≥
a
2
‖d‖2 + o(‖d‖2),
which implies that both sides are positive for small d.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 102 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.5
Let f ∈ C 1 be convex on the convex set Ω. If there is a point x∗ ∈ Ω such
that, for all y ∈ Ω, ∇f (x∗)(y − x) ≥ 0, then x∗ is a global minimum point
of f over Ω.
Proof.
Since y − x∗ is a feasible direction at x∗, the given condition is equivalent
to the first-order necessary condition. The proof is immediate, since by
Proposition 2.3, we have
f (y) ≥ f (x∗) +∇f (x∗)(y − x∗) ≥ f (x∗).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 103 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.5
Let f ∈ C 1 be convex on the convex set Ω. If there is a point x∗ ∈ Ω such
that, for all y ∈ Ω, ∇f (x∗)(y − x) ≥ 0, then x∗ is a global minimum point
of f over Ω.
Proof.
Since y − x∗ is a feasible direction at x∗, the given condition is equivalent
to the first-order necessary condition. The proof is immediate, since by
Proposition 2.3, we have
f (y) ≥ f (x∗) +∇f (x∗)(y − x∗) ≥ f (x∗).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 103 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Consider the problem
min f (x), s.t. x ∈ Ω. (7)
Consider the set Γ ⊂ Rn+1 = {(r , x) : f (x) ≤ r , x ∈ Rn}. Suppose that
x∗ ∈ Ω is the minimizing point with f ∗ = f (x∗) and construct a tubular
region B with cross section Ω extending vertically from −∞ up to f ∗. B is
also a convex set, and it overlaps the convex set Γ only at the boundary
point (f ∗,b∗) above x∗.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 104 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.6 (Separating Hyperplane Theorem)
Let B and C be convex sets with no common relative interior points, i.e.,
the only common points are boundary points. Then there is a hyperplane
separating B and C , that is, there is a nonzero vector a such that
sup
b∈B
atb ≤ inf
c∈C
atc.
According to Theorem 2.6, there is a hyperplane separating these two sets.
sr + λtx ≥ c , for all x ∈ Rn and f (x) ≤ r , (8)
sr + λtx ≤ c , for all x ∈ Ω and r ≤ f ∗. (9)
It follows that s 6= 0; for otherwise λ 6= 0 and then (8) would be violated
for some x ∈ Rn. It also follows that s ≥ 0 since otherwise (9) would be
violated by very negative values of r . As a consequence, we find s > 0 and
by appropriate scaling we may take s = 1.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 105 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.7 (Zero-order necessary conditions)
If x∗ solves (7) under the stated convexity conditions, then there is a
nonzero vector λ ∈ Rn such that x∗ is a solution to the two problems,
min f (x) + λtx, s.t. x ∈ Rn, (10)
max λtx, s.t. x ∈ Ω. (11)
Proof.
Problem (10) follows from (8) (with s = 1) and the fact that f (x) ≤ r for
r ≥ f (x). The value c is attained from above at (f ∗, x∗). Likewise
problem (11) follows from (9) and the fact that x∗ and the appropriate r
attain c from below.
Example 2.27 (Investigate what will happen to f ∈ C 1 in R1 over [0, 1])
let f ∈ C 1 on Rn, and let f have a minimum with respect to Ω at x∗. Let
d ∈ Rn be a feasible direction at x∗. By (10), we have ∇f (x∗)d ≥ 0.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 106 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.7 (Zero-order necessary conditions)
If x∗ solves (7) under the stated convexity conditions, then there is a
nonzero vector λ ∈ Rn such that x∗ is a solution to the two problems,
min f (x) + λtx, s.t. x ∈ Rn, (10)
max λtx, s.t. x ∈ Ω. (11)
Proof.
Problem (10) follows from (8) (with s = 1) and the fact that f (x) ≤ r for
r ≥ f (x). The value c is attained from above at (f ∗, x∗). Likewise
problem (11) follows from (9) and the fact that x∗ and the appropriate r
attain c from below.
Example 2.27 (Investigate what will happen to f ∈ C 1 in R1 over [0, 1])
let f ∈ C 1 on Rn, and let f have a minimum with respect to Ω at x∗. Let
d ∈ Rn be a feasible direction at x∗. By (10), we have ∇f (x∗)d ≥ 0.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 106 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.8 (Zero-order sufficiency conditions)
If there is a λ such that x∗ ∈ Ω solves the problems (10) and (11), then x∗
solves (7).
Proof.
Suppose x1 is any other point in Ω. Then from (10),
f (x1) + λ
tx1 ≥ f (x∗) + λtx∗.
This can be rewritten as
f (x1)− f (x∗) ≥ λtx∗ − λtx1.
By problem (11), the right hand side is greater than or equal to zero.
Hence f (x1)− f (x∗) ≥ 0, which establishes the result.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 107 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Proposition 2.8 (Zero-order sufficiency conditions)
If there is a λ such that x∗ ∈ Ω solves the problems (10) and (11), then x∗
solves (7).
Proof.
Suppose x1 is any other point in Ω. Then from (10),
f (x1) + λ
tx1 ≥ f (x∗) + λtx∗.
This can be rewritten as
f (x1)− f (x∗) ≥ λtx∗ − λtx1.
By problem (11), the right hand side is greater than or equal to zero.
Hence f (x1)− f (x∗) ≥ 0, which establishes the result.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 107 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Line Search Methods
FIBONACCI and Golden Section Search
Line Search by Curve Fitting: Newton’s Method, Method of False
Position, Cubic Fit, Quadratic Fit
Inaccurate Line Search: Percentage Test, Armijo’s Rule, Goldstein
Test, Wolfe Test, Backtracking
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 108 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
FIBONACCI Search
FIBONACCI method determines the minimum value of a function f over a
closed interval [c1, c2]. The only property that is assumed of f is unimodal.
After values are known at N points x1, x2, . . . , xN with
c1 = x0 ≤ x1 < x2 < . . . < xN−1 < xN ≤ xN+1 = c2,
the region of uncertainty is the interval [xk−1, xk+1] where xk is the
minimum point among the N.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 109 / 139
Xi CHEN
t1
Xi CHEN
Xi CHEN
Xi CHEN
t2
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Let d1 = c2 − c1 be the initial width of uncertainty, and dk be the width of
uncertainty after k measurements. If we have N measurements in total,
dk =
(
FN−k+1
FN
)
d1,
where the integers Fk are members of the Fibonacci sequence generated
by the recurrence relation FN = FN−1 + FN−2 with F0 = F1 = 1.
Recursive Structure
Ln−1 = 2Ln − �;
Ln−2 = Ln−1 + Ln = 3Ln − �;
Ln−3 = Ln−2 + Ln−1 = 5Ln − 2�;
. . .
Ln−k = Ln−(k−1) + Ln−(k−2) = Fk+1Ln − Fk−1�.
To reach certain pre-specified accuracy δ, we need to satisfy Fk ≥ 1/δ.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 110 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
FIBONACCI Search Algorithm (Minimization)
1 Determine k according to δ, and a0, b0. Compute t1 and t2 by
t1 = b0 −
Fk−1
Fk
(b0 − a0), t2 = a0 +
Fk−1
Fk
(b0 − a0).
2 If f (t1) < f (t2), set a1 = a0, b1 = t2, t2 = t1 and compute t1 by
t1 = b1 −
Fk−2
Fk−1
(b1 − a1).
If f (t1) ≥ f (t2), set b1 = b0, a1 = t1, t1 = t2 and compute t2 by
t2 = a1 +
Fk−2
Fk−1
(b1 − a1).
3 . . .
4
t1 = bn −
F1
F2
(bn − an), or t2 = an +
F1
F2
(bn − an).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 111 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
The sequence of measurement points is determined in accordance with the
assumption that each measurement is of lower value than its predecessors.
Note that the procedure always calls for the last two measurements to be
made at the midpoint of the semifinal interval of uncertainty.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 112 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Golden Section Search
Suppose xn+2 = c1xn+1 + c2xn. If we can transform it into
xn+2 − rxn+1 = s(xn+1 − rxn), then xn+2 = (r + s)xn+1 − rsxn. By
comparison, the characteristic equation can be derived by cancelling s or r .
As for the Fibonacci difference equation,
τ2 − τ − 1 = 0⇒ τ1 =
1 +
√
5
2
, τ2 =
1−
√
5
2
,
where τ1 is known as the golden section ratio. The solution to the
Fibonacci difference equation FN = FN−1 + FN−2 is of the form
FN = Aτ
N
1 + Bτ
N
2 ⇒ lim
N→∞
FN−1
FN
=
1
τ1
' 0.618.
As a consequence, we have
dk =
(
1
τ1
)k−1
d1,
dk+1
dk
=
1
τ1
' 0.618.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 113 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Golden Section Search Algorithm (Minimization)
1 Initialize a, b, and �. Let λ = 0.618. Compute x1 and x2 by
x1 = a + (1− λ)(b − a), x2 = a + λ(b − a).
and f1 = f (x1), f2 = f (x2).
2 If f (x1) < f (x2), set b = x2, x2 = x1, f2 = f1 and compute x1 by
x1 = a + (1− λ)(b − a).
and f1 = f (x1).
If f (x1) ≥ f (x2), set a = x1, x1 = x2, f1 = f2 and compute x2 by
x2 = a + λ(b − a).
and f2 = f (x2).
3 If |a− b| < �, then x∗ = 0.5(a + b), f ∗ = f (x∗); else go to Step 2.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 114 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Example 2.28
By using Fibonacci Search, maximize the function
f (x) = −|2− x | − |5− 4x | − |8− 9x |
over the interval [0, 3]. The user specified tolerance level is � = 0.1, and
the first two experimental endpoints are x1 = 1.147 and x2 = 1.853.
Iteration x1 x2 f (x1) f (x2) Interval
2 0.7059 1.1471 -3.5882 -11.2353 [0.0000, 1.8529]
3 1.1471 1.4118 -5.1176 -3.5882 [0.7059, 1.8529]
4 0.9706 1.1471 -3.5882 -5.9412 [0.7059, 1.4118]
5 0.8824 0.9706 -2.8824 -3.5882 [0.7059, 1.1471]
6 0.7941 0.8824 -2.6471 -2.8824 [0.7059, 0.9706]
7 0.8824 0.8824 -3.8824 -2.6471 [0.7941, 0.9706]
Example 2.29
Solve the previous problem by using Golden Section Search.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 115 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Newton’s Method
In most problems, however, it can be safely assumed that the function
being searched, as well as being unimodal, possesses a certain degree of
smoothness, and one might, therefore, expect that more efficient search
techniques exploiting this smoothness can be devised.
Suppose that the function f of a single variable x is to be minimized, and
it is possible to evaluate f (xk), f
′(xk), f
′′(xk).
q(x) = f (xk) + f
′(xk)(x − xk) +
1
2
f ′′(xk)(x − xk)2,
q′(x) = 0⇒ xk+1 = xk −
f ′(xk)
f ′′(xk)
.
Letting g(x) ≡ f ′(x), we have xk+1 = xk −
g(xk)
g ′(xk)
.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 116 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
The method can more simply be viewed as a technique for iteratively
solving equations of the form g(x) = 0 by letting g(x) ≡ f ′(x).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 117 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Method of False Position
Newton’s method for minimization is based on fitting a quadratic on the
basis of information at a single point. However, by using more points, e.g.,
f (xk), f
′(xk), f
′(xk−1), less information is required at each of them to fit
the quadratic
q(x) = f (xk) + f
′(xk)(x − xk) +
f ′(xk−1)− f ′(xk)
xk−1 − xk
·
(x − xk)2
2
,
xk+1 = xk − f ′(xk)
[
xk−1 − xk
f ′(xk−1)− f ′(xk)
]
.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 118 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Since this method does not depend on values of f directly, it can be
regarded as a method for solving f ′(x) ≡ g(x) = 0. Viewed in this way the
method takes the form
xk+1 = xk − g(xk)
[
xk − xk−1
g(xk)− g(xk−1)
]
.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 119 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Definition 2.6
Let the sequence {rk}∞k=0 converge to r
∗. The order of convergence of
{rk} is defined as the supremum of the nonnegative numbers p satisfying
0 ≤ lim
k→∞
|rk+1 − r∗|
|rk − r∗|p
<∞.
Definition 2.7
If the sequence {rk}∞k=0 converge to r
∗ in such a way that
lim
k→∞
|rk+1 − r∗|
|rk − r∗|
= β < 1,
the sequence is said to converge linearly to r∗ with convergence ratio (or
rate) β.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 120 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
When comparing the relative effectiveness of two competing algorithms,
both of which produce linearly convergent sequences, the comparison is
based on their corresponding convergence ratios:
The smaller the ratio the faster the rate.
The ultimate case where β = 0 is referred to as superlinear convergence.
Convergence of any order greater than one, i.e., p > 1, is superlinear, but
it is also possible for superlinear convergence to correspond to unity order.
Example 2.30
The sequence rk = 1/k converges to zero. The convergence is of order
one but it is not linear, since limk→∞(rk+1/rk) = 1, that is, β = 1.
Example 2.31
The sequence rk = (1/k)
k is of order one, since rk+1/(rk)
p →∞ for
p > 1. But limk→∞ rk+1/rk = 0 and hence this is superlinear convergence.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 121 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.7
Let the function g have a continuous second derivative, and let x∗ satisfy
g(x∗) = 0 and g ′(x∗) 6= 0. Then, provided x0 is sufficiently close to x∗,
the sequence {xk} generated by Newton’s method converges to x∗ with an
order of convergence at least two.
Proof.
Since g(x∗) = 0, we have
xk+1 − x∗ = xk − x∗ −
g(xk)− g(x∗)
g ′(xk)
= −
g(xk)− g(x∗) + g ′(xk)(x∗ − xk)
g ′(xk)
=
1
2
g ′′(ξ)
g ′(xk)
(xk − x∗)2
for some ξ between x∗ and xk . Therefore, if started close enough to the
solution, the method will converge to x∗ with an order of convergence at
least two.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 122 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.7
Let the function g have a continuous second derivative, and let x∗ satisfy
g(x∗) = 0 and g ′(x∗) 6= 0. Then, provided x0 is sufficiently close to x∗,
the sequence {xk} generated by Newton’s method converges to x∗ with an
order of convergence at least two.
Proof.
Since g(x∗) = 0, we have
xk+1 − x∗ = xk − x∗ −
g(xk)− g(x∗)
g ′(xk)
= −
g(xk)− g(x∗) + g ′(xk)(x∗ − xk)
g ′(xk)
=
1
2
g ′′(ξ)
g ′(xk)
(xk − x∗)2
for some ξ between x∗ and xk . Therefore, if started close enough to the
solution, the method will converge to x∗ with an order of convergence at
least two.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 122 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
The derivative of f from M0 = (x0, y0, z0)
t along the direction ` is
∂f (M0)
∂`
=
∂f
∂x
· cosα +
∂f
∂y
· cosβ +
∂f
∂z
· cos γ.
Proof.
Let M = (x , y , z)t be an arbitrary point along `, and let t = M0M. Then
we have x − x0 = t cosα, y − y0 = t cosβ, and z − z0 = t cos γ.
∂f (M0)
∂`
= lim
M→M0
f (M)− f (M0)
M0M
= lim
t→0
φ(t)− φ(0)
t
= φ′(0).
On the other hand, we have
φ′(t) =
∂f
∂x
·
dx
dt
+
∂f
∂y
·
dy
dt
+
∂f
∂z
·
dz
dt
=
∂f
∂x
· cosα +
∂f
∂y
· cosβ +
∂f
∂z
· cos γ.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 123 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
The derivative of f from M0 = (x0, y0, z0)
t along the direction ` is
∂f (M0)
∂`
=
∂f
∂x
· cosα +
∂f
∂y
· cosβ +
∂f
∂z
· cos γ.
Proof.
Let M = (x , y , z)t be an arbitrary point along `, and let t = M0M. Then
we have x − x0 = t cosα, y − y0 = t cosβ, and z − z0 = t cos γ.
∂f (M0)
∂`
= lim
M→M0
f (M)− f (M0)
M0M
= lim
t→0
φ(t)− φ(0)
t
= φ′(0).
On the other hand, we have
φ′(t) =
∂f
∂x
·
dx
dt
+
∂f
∂y
·
dy
dt
+
∂f
∂z
·
dz
dt
=
∂f
∂x
· cosα +
∂f
∂y
· cosβ +
∂f
∂z
· cos γ.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 123 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Let
a =
∂f (M0)
∂x
, b =
∂f (M0)
∂y
, c =
∂f (M0)
∂z
.
Then we have
∂f
∂`
=
∂f
∂x
· cosα +
∂f
∂y
· cosβ +
∂f
∂z
· cos γ
= a · cosα + b · cosβ + c · cos γ
=
√
a2 + b2 + c2(cosα1 · cosα + cosβ1 · cosβ + cos γ1 · cos γ)
=
√
a2 + b2 + c2 cos(`1, `) direction
≤
√(
∂f
∂x
)2
+
(
∂f
∂y
)2
+
(
∂f
∂z
)2
length
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 124 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Steepest Descent Method (Gradient Method)
By Taylor’s theorem, if f ∈ C 1 in a region containing the line segment
[xk , xk+1], then there is an α, 0 ≤ α ≤ 1, such that
f (xk+1) = f (xk) +∇f (xk)(xk+1 − xk)
= f (xk) +∇f (xk)dk .
Apparently, if dk = −∇f (xk)t , then we have
f (xk+1) = f (xk)− ‖∇f (xk)t‖2,
which can guarantee the descent of a minimizing problem provided
∇f (xk)t 6= 0, or practically, ‖∇f (xk)t‖2 < �. By setting dk = −∇f (xk)t ,
we need to choose the step-size αk such that
αk = arg min
α
f
(
xk − α∇f (xk)t
)
.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 125 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
The Quadratic Case
Consider
f (x) =
1
2
xtHx− xtb and E (x) =
1
2
(x− x∗)tH(x− x∗), (12)
where H is a positive definite symmetric n × n matrix.
The steepest descent method can be expressed as xk+1 = xk − αkgk ,
where gk = g(xk) = ∇f (xk)t = Hxk − b, and αk minimizes f (xk − αgk).
By (12), we have
f (xk − αgk) =
1
2
(xk − αgk)tH(xk − αgk)− (xk − αgk)tb,
and thus
αk =
gtkgk
gtkHgk
⇒ xk+1 = xk −
(
gtkgk
gtkHgk
)
gk .
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 126 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Example 2.32
The function f and the steepest descent process can be illustrated as in
the following figure by showing contours of constant values of f and a
typical sequence developed by the process. The contours of f are
n-dimensional ellipsoids with axes in the directions of the n-mutually
orthogonal eigenvectors of H.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 127 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Lemma 2.1
The iterative process of steepest descent method satisfies
E (xk+1) =
{
1−
(gtkgk)
2
(gtkHgk)(g
t
kH
−1gk)
}
E (xk).
Proof.
Setting yk = xk − x∗, gk = Hxk − b = Hxk −Hx∗ = Hyk . For E (yk), we
have
E (xk)− E (xk+1)
E (xk)
=
2αkg
t
kHyk − α
2
kg
t
kHgk
ytkHyk
=
(gtkgk)
2
(gtkHgk)(g
t
kH
−1gk)
.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 128 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Lemma 2.1
The iterative process of steepest descent method satisfies
E (xk+1) =
{
1−
(gtkgk)
2
(gtkHgk)(g
t
kH
−1gk)
}
E (xk).
Proof.
Setting yk = xk − x∗, gk = Hxk − b = Hxk −Hx∗ = Hyk . For E (yk), we
have
E (xk)− E (xk+1)
E (xk)
=
2αkg
t
kHyk − α
2
kg
t
kHgk
ytkHyk
=
(gtkgk)
2
(gtkHgk)(g
t
kH
−1gk)
.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 128 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Lemma 2.2 (Kantorovich inequality)
Let H be a positive definite symmetric n × n matrix. For any vector x
there holds
(xtx)2
(xtHx)(xtH−1x)
≥
4aA
(a + A)2
where a and A are, respectively, the smallest and largest eigenvalues of H.
Proof.
Let the eigenvalues of H satisfy a = λ1 ≤ λ2 ≤ . . . ≤ λn = A. By an
appropriate change of coordinates the matrix H becomes diagonal with
diagonal λ1, λ2, . . . , λn. In this coordinate system we have
(xtx)2
(xtHx)(xtH−1x)
=
(∑n
i=1 x
2
i
)2(∑n
i=1 λix
2
i
) [∑n
i=1(x
2
i /λi )
] ,
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 129 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Lemma 2.2 (Kantorovich inequality)
Let H be a positive definite symmetric n × n matrix. For any vector x
there holds
(xtx)2
(xtHx)(xtH−1x)
≥
4aA
(a + A)2
where a and A are, respectively, the smallest and largest eigenvalues of H.
Proof.
Let the eigenvalues of H satisfy a = λ1 ≤ λ2 ≤ . . . ≤ λn = A. By an
appropriate change of coordinates the matrix H becomes diagonal with
diagonal λ1, λ2, . . . , λn. In this coordinate system we have
(xtx)2
(xtHx)(xtH−1x)
=
(∑n
i=1 x
2
i
)2(∑n
i=1 λix
2
i
) [∑n
i=1(x
2
i /λi )
] ,
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 129 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
continued.
which, taking into account the convex combination, can be written as
(xtx)2
(xtHx)(xtH−1x)
=
1/(
∑n
i=1 ξiλi )∑n
i=1(ξi/λi )
≡
φ(ξ)
ψ(ξ)
, where ξi =
x2i∑n
i=1 x
2
i
.
The minimum value of this ratio is achieved for some λ = ξ1λ1 + ξnλn
with ξ1 + ξn = 1. Using the relation
ξ1
λ1
+
ξn
λn
=
λ1 + λn − ξ1λ1 − ξnλn
λ1λn
,
an appropriate bound is
φ(ξ)
ψ(ξ)
≥ lim
λ1≤λ≤λn
1/λ
(λ1 + λn − λ)/(λ1λn)
≥
4λ1λn
(λ1 + λn)2
,
where the minimum is achieved at λ = (λ1 + λn)/2.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 130 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
continued.
φ(ξ)
ψ(ξ)
=
1/(
∑n
i=1 ξiλi )∑n
i=1(ξi/λi )
The curve represents the function 1/λ. Since
∑n
i=1 ξiλi is a point between
λ1 and λn, the value of φ(ξ) is a point on the curve. On the other hand,
the value of ψ(ξ) is a convex combination of points on the curve and its
value corresponds to a point in the shaded region. For the same vector ξ,
both functions are represented by points on the same vertical line.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 131 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.8 (Steepest descent method–quadratic case)
For any x0 ∈ Rn, the steepest descent method converges to the unique
minimum point x∗ of f . Furthermore, there holds at every step k
E (xk+1) ≤
(
A− a
A + a
)2
E (xk).
Proof.
By Lemma 2.1 and Lemma 2.2, we have
E (xk+1) =
{
1−
(gtkgk)
2
(gtkHgk)(g
t
kH
−1gk)
}
E (xk)
≤
[
1−
4aA
(A + a)2
]
E (xk) =
(
A− a
A + a
)2
E (xk).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 132 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.8 (Steepest descent method–quadratic case)
For any x0 ∈ Rn, the steepest descent method converges to the unique
minimum point x∗ of f . Furthermore, there holds at every step k
E (xk+1) ≤
(
A− a
A + a
)2
E (xk).
Proof.
By Lemma 2.1 and Lemma 2.2, we have
E (xk+1) =
{
1−
(gtkgk)
2
(gtkHgk)(g
t
kH
−1gk)
}
E (xk)
≤
[
1−
4aA
(A + a)2
]
E (xk) =
(
A− a
A + a
)2
E (xk).
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 132 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
In terms of convergence rate, Theorem 2.8 states that regarding E (x), the
steepest descent method converges linearly with a ratio no greater than(
A− a
A + a
)2
.
The convergence rate actually depends only on the ratio r = A/a. Thus
the convergence ratio is (
A− a
A + a
)2
=
(
r − 1
r + 1
)2
,
which clearly shows that convergence is slowed as r increases. The ratio r ,
which is the single number associated with the matrix H that characterizes
convergence, is often called the condition number of the matrix.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 133 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Example 2.33
Consider
H =
0.78 −0.02 −0.12 −0.14
−0.02 0.86 −0.04 0.06
−0.12 −0.04 0.72 −0.08
−0.14 0.06 −0.08 0.74
, b =
0.76
0.08
1.12
0.68
.
Starting with x0 = (0, 0, 0, 0)
t , show the solution sequence generated by
the steepest descent method converges to
x∗ = (1.534965, 0.1220097, 1.975156, 1.412954)t , f (x∗) = −2.1746595.
For this positive definite matrix H, we have a = 0.52, A = 0.94 and hence
r = 1.8. This is a very favorable condition number and leads to the
convergence ratio is 0.081. Thus each iteration will reduce the error in the
objective by more than a factor of ten; or, equivalently, each iteration will
add about one more digit of accuracy.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 134 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Steepest descent method–nonquadratic case
To establish estimates of the progress of the gradient method when the
Hessian matrix is always positive definite, assume that the Hessian matrix
is bounded above and below as aI ≤ F(x) ≤ AI.
Given a point xk , we have for any α,
f (xk − αg(xk)) ≤ f (xk)− αg(xk)tg(xk) +
Aα2
2
g(xk)
tg(xk).
Minimizing both sides separately with respect to α the inequality will hold
for the two minima, that is,
f (xk+1) ≤ f (xk)−
1
2A
‖g(xk)‖2,
and we thus have
f (xk+1)− f ∗ ≤ f (xk)− f ∗ −
1
2A
‖g(xk)‖2. (13)
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 135 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
In a similar way, for any x there holds
f (x) ≥ f (xk) + g(xk)t(x− xk) +
a
2
‖x− xk‖2.
Again we can minimize both sides separately and have
f ∗ ≥ f (xk)−
1
2a
‖g(xk)‖2 ⇒ −‖g(xk)‖2 ≤ 2a[f ∗ − f (xk)]. (14)
By (13) and (14), we have
f (xk+1)− f ∗
f (xk)− f ∗
≤ 1−
a
A
,
which shows that the gradient method makes progress even when the
starting point is not close to the optimal solution.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 136 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Example 2.34
To minimize a function f , consider solving the equations ∇f (x) = 0 that
represent the necessary conditions. We could apply steepest descent to the
function h(x) = ‖∇f (x)‖2. For simplicity, consider the quadratic case,
f (x) =
1
2
xtQx− btx, h(x) = xtQ2x− 2xtQb + btb.
The rate of convergence of steepest descent applied to h(x) will be
governed by the eigenvalues of the matrix Q2. By(
r − 1
r + 1
)2
=
(
r2 − 1
r2 + 1
)2
'
(
1−
1
r2
)4
,(
r − 1
r + 1
)2
'
(
1−
1
r
)4
,
(
1−
1
r2
)r
' 1−
1
r
,
it takes about r steps of the new method to equal one step of ordinary
steepest descent.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 137 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Newton’s Method
The function f being minimized is approximated locally by a quadratic
function, and this approximate function is minimized exactly. Thus near
xk , we can approximate f by the truncated Taylor series,
f (x) ' f (xk) +∇f (xk)(x− xk) +
1
2
(x− xk)tF(xk)(x− xk).
The right-hand side is minimized at
xk+1 = xk − [F(xk)]−1∇f (xk)t ,
and this equation is the pure form of Newton’s method.
In view of the second-order sufficiency conditions for a minimum point, we
assume that at a relative minimum point, x∗, the Hessian matrix, F(x∗), is
positive definite.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 138 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.9
Let f ∈ C 3 on Rn, and assume that at the local minimum point x∗, the
Hessian F(x∗) is positive definite. Then if started sufficiently close to x∗,
the points generated by Newton’s method converge to x∗. The order of
convergence is at least two.
Proof.
There are ρ > 0, β1 > 0, β2 > 0 such that ∀x with ‖x− x∗‖ < ρ, there
holds ‖F(x)−1‖ < β1, ‖∇f (x∗)t −∇f (x)t − F(x)(x∗ − x)‖ ≤ β2‖x∗ − x‖2.
Suppose xk is selected with β1β2‖x∗ − xk‖ < 1 and ‖x∗ − xk‖ < ρ. Then
‖xk+1 − x∗‖ =
∥∥xk − x∗ − [F(xk)]−1∇f (xk)t∥∥
=
∥∥[F(xk)]−1 [∇f (x∗)t −∇f (xk)t − F(xk)(x∗ − xk)]∥∥
≤
∥∥[F(xk)]−1∥∥ · β2‖xk − x∗‖2≤ β1β2‖xk − x∗‖2 < ‖xk − x∗‖.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 139 / 139
���`̀̀zzz¯̄̄KKK¦¦¦))) ���ggg555yyy
Theorem 2.9
Let f ∈ C 3 on Rn, and assume that at the local minimum point x∗, the
Hessian F(x∗) is positive definite. Then if started sufficiently close to x∗,
the points generated by Newton’s method converge to x∗. The order of
convergence is at least two.
Proof.
There are ρ > 0, β1 > 0, β2 > 0 such that ∀x with ‖x− x∗‖ < ρ, there
holds ‖F(x)−1‖ < β1, ‖∇f (x∗)t −∇f (x)t − F(x)(x∗ − x)‖ ≤ β2‖x∗ − x‖2.
Suppose xk is selected with β1β2‖x∗ − xk‖ < 1 and ‖x∗ − xk‖ < ρ. Then
‖xk+1 − x∗‖ =
∥∥xk − x∗ − [F(xk)]−1∇f (xk)t∥∥
=
∥∥[F(xk)]−1 [∇f (x∗)t −∇f (xk)t − F(xk)(x∗ − xk)]∥∥
≤
∥∥[F(xk)]−1∥∥ · β2‖xk − x∗‖2≤ β1β2‖xk − x∗‖2 < ‖xk − x∗‖.
Xi Chen ( .cn) MATLAB 333²²²LLL���+++nnnïïïÄÄÄ¥¥¥���AAA^̂̂ 139 / 139
MATLAB »ù´¡
¾ØÕóÓëÊý×éÔËËã
³£ÓÃÊýѧº¯Êý
·ûºÅÔËËã
MATLAB ±à³Ì
¶þάÓëÈýά»æͼ
×îÓÅ»¯ÎÊÌâÇó½â
ÏßÐԹ滮
ÕûÊý¹æ»®
¶þ´Î¹æ»®