多变量优化问题应该是在高阶导问题(看这里优化方法理论合集(2)——高阶导问题)之后更新的,由于笔者失误,特在此补充。所幸不影响之前优化方法的理解与学习。
1. 问题描述
多变量问题描述的是关于质点在 n n n维空间中坐标的问题。
给出如下条件:
- 时间区间
t ∈ [ t 0 , t k ] t \in \left[ t_0, t_k \right] t∈[t0,tk] - 系统有
2
n
2n
2n个边界条件:
S t a r t = { x 1 ( t 0 ) = x 10 x 2 ( t 0 ) = x 20 ⋮ x n ( t 0 ) = x n 0 Start = \begin{cases} x_1(t_0) = x_{10} \\ x_2(t_0) = x_{20} \\ \vdots \\ x_n(t_0) = x_{n0} \end{cases} Start=⎩⎪⎪⎪⎪⎨⎪⎪⎪⎪⎧x1(t0)=x10x2(t0)=x20⋮xn(t0)=xn0 E n d = { x 1 ( t k ) = x 1 k x 2 ( t k ) = x 2 k ⋮ x n ( t k ) = x n k End = \begin{cases} x_1(t_k) = x_{1k} \\ x_2(t_k) = x_{2k} \\ \vdots \\ x_n(t_k) = x_{nk} \end{cases} End=⎩⎪⎪⎪⎪⎨⎪⎪⎪⎪⎧x1(tk)=x1kx2(tk)=x2k⋮xn(tk)=xnk - 性能指标(拉格朗日形式):
J = ∫ t 0 t k F ( t , x 1 , x ˙ 1 , x 2 , x ˙ 2 , ⋯ , x n , x ˙ n ) d t → e x t r . J = \int _{t_0} ^{t_k} F \left( t, x_1, \dot x_1, x_2, \dot x_2, \cdots, x_n, \dot x_n \right) dt \rightarrow extr. J=∫t0tkF(t,x1,x˙1,x2,x˙2,⋯,xn,x˙n)dt→extr.
2. 解题步骤
写出解题的必要条件:
{
F
x
1
−
d
d
t
F
x
˙
1
=
0
F
x
2
−
d
d
t
F
x
˙
2
=
0
⋮
F
x
n
−
d
d
t
F
x
˙
n
=
0
(1)
\begin{cases} F_{x_1} - \frac{d}{dt} F _{\dot x_1} = 0 \\ F_{x_2} - \frac{d}{dt} F _{\dot x_2} = 0 \\ \vdots \\ F_{x_n} - \frac{d}{dt} F _{\dot x_n} = 0 \tag{1} \end{cases}
⎩⎪⎪⎪⎪⎨⎪⎪⎪⎪⎧Fx1−dtdFx˙1=0Fx2−dtdFx˙2=0⋮Fxn−dtdFx˙n=0(1)即对每个
F
i
F_i
Fi列出欧拉方程。
与之前所提的不同,极值存在的充分条件不再是勒让德条件,而是盖西条件:
H
=
∣
∂
2
F
∂
x
˙
1
x
˙
1
∂
2
F
∂
x
˙
1
x
˙
2
⋯
∂
2
F
∂
x
˙
1
x
˙
n
∂
2
F
∂
x
˙
2
x
˙
1
∂
2
F
∂
x
˙
2
x
˙
2
⋯
∂
2
F
∂
x
˙
2
x
˙
n
⋮
⋮
⋯
⋮
∂
2
F
∂
x
˙
n
x
˙
1
∂
2
F
∂
x
˙
n
x
˙
2
⋯
∂
2
F
∂
x
˙
n
x
˙
n
∣
(2)
H = \left| \begin{matrix} \frac{\partial ^2 F}{\partial \dot x_1 \dot x _1} & \frac{\partial ^2 F}{\partial \dot x_1 \dot x _2} & \cdots & \frac{\partial ^2 F}{\partial \dot x_1 \dot x _n} \\ \frac{\partial ^2 F}{\partial \dot x_2 \dot x _1} & \frac{\partial ^2 F}{\partial \dot x_2 \dot x _2} & \cdots & \frac{\partial ^2 F}{\partial \dot x_2 \dot x _n} \\ \vdots & \vdots & \cdots & \vdots \\ \frac{\partial ^2 F}{\partial \dot x_n \dot x _1} & \frac{\partial ^2 F}{\partial \dot x_n \dot x _2} & \cdots & \frac{\partial ^2 F}{\partial \dot x_n \dot x _n} \\ \end{matrix} \right| \tag{2}
H=∣∣∣∣∣∣∣∣∣∣∂x˙1x˙1∂2F∂x˙2x˙1∂2F⋮∂x˙nx˙1∂2F∂x˙1x˙2∂2F∂x˙2x˙2∂2F⋮∂x˙nx˙2∂2F⋯⋯⋯⋯∂x˙1x˙n∂2F∂x˙2x˙n∂2F⋮∂x˙nx˙n∂2F∣∣∣∣∣∣∣∣∣∣(2)同时有
{
若
H
>
0
,
则
J
→
m
i
n
若
H
<
0
,
则
J
→
m
a
x
\begin{cases} 若H > 0,则J \rightarrow min \\ 若H < 0,则J \rightarrow max \end{cases}
{若H>0,则J→min若H<0,则J→max
3. 例题
给出如下条件:
- 时间区间
t ∈ [ 0 , T ] t \in \left[ 0, T \right] t∈[0,T] - 边界条件
S t a r t = { x 1 ( 0 ) = x 10 x 2 ( 0 ) = x 20 Start = \begin{cases} x_1(0) = x_{10} \\ x_2(0) = x_{20} \end{cases} Start={x1(0)=x10x2(0)=x20 E n d = { x 1 ( T ) = x 1 k x 2 ( T ) = x 2 k End = \begin{cases} x_1(T) = x_{1k} \\ x_2(T) = x_{2k} \end{cases} End={x1(T)=x1kx2(T)=x2k - 性能指标
J = ∫ 0 T ( x ˙ 1 2 + x ˙ 2 2 + 2 x 1 x 2 ) d t → e x t r . J = \int _0 ^T \left( \dot x_1^2 + \dot x_2^2 + 2x_1 x_2 \right) dt \rightarrow extr. J=∫0T(x˙12+x˙22+2x1x2)dt→extr.
解: □ \square \quad □把 F = x ˙ 1 2 + x ˙ 2 2 + 2 x 1 x 2 F = \dot x_1^2 + \dot x_2^2 + 2x_1 x_2 F=x˙12+x˙22+2x1x2代入欧拉方程组(1)有
{ F x 1 − d d t F x ˙ 1 = 0 ⟹ 2 x 2 − 2 x ¨ 1 = 0 F x 2 − d d t F x ˙ 2 = 0 ⟹ 2 x 1 − 2 x ¨ 2 = 0 \begin{cases} F_{x_1} - \frac{d}{dt} F_{\dot x_1} = 0 \Longrightarrow 2x_2 - 2 \ddot x_1 = 0 \\ F_{x_2} - \frac{d}{dt} F_{\dot x_2} = 0 \Longrightarrow 2x_1 - 2 \ddot x_2 = 0 \end{cases} {Fx1−dtdFx˙1=0⟹2x2−2x¨1=0Fx2−dtdFx˙2=0⟹2x1−2x¨2=0解出
x 2 ( t ) = C 1 e t + C 2 e − t + C 3 e i t + C 4 e − i t , x 1 ( t ) = x ¨ 2 ( t ) = C 1 e t + C 2 e − t − C 3 e i t + C 4 e − i t x_2(t) = C_1 e^t + C_2 e^{-t} + C_3 e^{it} + C_4 e^{-it}, \\ x_1 (t) = \ddot x_2 (t) = C_1 e^t + C_2 e^{-t} - C_3 e^{it} + C_4 e^{-it} x2(t)=C1et+C2e−t+C3eit+C4e−it,x1(t)=x¨2(t)=C1et+C2e−t−C3eit+C4e−it代入初始条件值即可算出 C i C_i Ci。
又因为
H
=
∣
∂
2
F
∂
x
˙
1
∂
x
˙
1
∂
2
F
∂
x
˙
1
∂
x
˙
2
∂
2
F
∂
x
˙
2
∂
x
˙
1
∂
2
F
∂
x
˙
2
∂
x
˙
2
∣
=
∣
2
0
0
2
∣
=
4
>
0
H = \left| \begin{matrix} \frac{\partial ^2 F}{\partial \dot x_1 \partial \dot x_1} & \frac{\partial ^2 F}{\partial \dot x_1 \partial \dot x_2} \\ \frac{\partial ^2 F}{\partial \dot x_2 \partial \dot x_1} & \frac{\partial ^2 F}{\partial \dot x_2 \partial \dot x_2} \end{matrix} \right| = \left| \begin{matrix} 2 & 0 \\ 0 & 2 \end{matrix} \right| = 4 > 0
H=∣∣∣∣∣∂x˙1∂x˙1∂2F∂x˙2∂x˙1∂2F∂x˙1∂x˙2∂2F∂x˙2∂x˙2∂2F∣∣∣∣∣=∣∣∣∣2002∣∣∣∣=4>0所以
J
J
J有最小值。