One dimensional optimization
Download
Report
Transcript One dimensional optimization
Optimality Conditions for
Unconstrained optimization
• One dimensional optimization
– Necessary and sufficient conditions
• Multidimensional optimization
– Classification of stationary points
– Necssary and sufficient conditions for local
optima.
• Convexity and global optimality
One dimensional optimization
• We are accustomed to think that if f(x) has
a minimum then f’(x)=0 but….
5
|x-5|
4.5
0.2(x-5)2
4
3.5
3
2.5
2
1.5
1
0.5
0
0
1
2
3
4
5
6
7
8
9
10
1D Optimization jargon
• A point with zero
derivative is a
stationary point.
• x=5, Can be
25
20
15
10
5
0
-5
a minimum
2
y x 5
A maximum
y 10 x x 2
3
An inflection point y 0.2 x 5
-10
-15
-20
-25
0
1
2
3
4
5
6
7
8
9
10
Optimality criteria for smooth 1D
functions at point x*
• f’(x*)=0 is the condition for stationarity and
a necessary condition for a minimum or a
maximum.
• f“(x*)>0 is sufficient for a minimum
• f“(x*)<0 is sufficient for a maximum
• With f”(x*)=0 needs information from
higher derivatives.
• Example?
Problems 1D
• Classify the stationary points of the following functions
from the optimality conditions, then check by plotting
them
1.
2.
3.
4.
2x3+3x2
3x4+4x3-12x2
x5
f=x4+4x3+6x2+4x
• Answer true or false:
– A function can have a negative value at its maximum point.
– If a constant is added to a function, the location of its minimum
point can change.
– If the curvature of a function is negative at a stationary point,
then the point is a maximum.
Taylor series expansion in n
dimensions
• Expanding f ( x1 , xn ) about a candidate
minimum x*
1 n
* f
f (x) f (x*) xi xi (x*)
xi
2 j 1
i 1
n
2
f
*
*
x
x
x
x
i i j j x x (x*)
i 1
i
j
n
1 T
f (x*) x f (x*) x H (x*)x
2
T
if
f
0 choose xi xi* of opposite sign and other x j x*j 0
xi
So must have f 0
• This is the condition for stationarity
Conditions for minimum
1 T
f (x) f (x*) x f (x*) x H (x*)x
2
T
• Sufficient condition for a minimum is that
xT H (x*)x 0 for all x
• That is, the matrix of second derivatives
(Hessian) is positive definite
• Simplest way to check positive definiteness is
eigenvalues: All eigenvalues need to be positive
• Necessary conditions matrix is positive-semi
definite, all eigenvalues non-negative
Types of stationary points
•
•
•
•
•
Positive definite: Minimum
Positive semi-definite: possibly minimum
Indefinite: Saddle point
Negative semi-definite: possibly maximum
Negative definite: maximum
Example
f x x1 x2 x
2
1
2
2
Stationary point:
2 x1 x2
f
x1 2 x2
x1 x2 0
12
10
Hessian matrix
8
2 f
2 f
x 2
x
x
2 1
1
1
2
H
2
2
f
1
2
f
2
x2
x1x2
Eigenvalues: 1,2 1,3 minimum
Change to f x 3 x1 x2 x
2
1
2
2
Eigenvalues: 1,2 1,5 saddle point
6
4
2
0
2
1
2
1
0
0
-1
-1
-2
x2
-2
x1
20
15
10
5
0
-5
2
1
2
1
0
0
-1
x2
-1
-2
-2
x1
Problems n-dimensional
• Find the stationary points of the following
functions and classify them:
1. x12 4 x1 x2 2 x1 x3 7 x22 6 x2 x3 5 x32
2. x 2 x2 x3 x 4 x
2
1
2
2
3.40 x1 x12 x2 x22 / x1
2
3
Global optimization
• The function x+sin(2x)
12
10
8
6
4
2
0
0
1
2
3
4
5
6
7
8
9
10
Convex function
• A straight line connecting two points will
not dip below the function graph.
f x1 (1 )x 2 f (x1 ) (1 ) f x 2
• Convex function will have a single
minimum.
10
5
0
Sufficient condition: Positive
semi-definite Hessian
everywhere.
-5
-10
-15
-20
-25
0
1
2
3
4
5
6
7
8
9
10
Problems convexity
• Check for convexity the following
functions. If the function is not convex
everywhere, check its domain of
convexity.
1.
2.
3 x12 2 x1 x2 2 x22 8
5 x1 x12 x22 4 x22 / x1
3. x13 12 x1 x22 2 x22 5 x12
Reciprocal approximation
• Reciprocal approximation (linear in one over the
variables) is desirable in many cases because it
captures decreasing returns behavior.
n
f
f
(
x
)
f
(
x
)
(
x
x
)
• Linear approximation L
0
i
0i
x
i 1
i x
• Reciprocal approximation
0
f
f L (y ) f (y 0 ) ( yi y0i )
yi 1/ xi
i 1
yi y 0
m
n
x0i
f R (x) f (x 0 ) ( xi x0i )
xi
i 1
f
xi x0
Conservative-convex
approximation
• At times we benefit from conservative
approximations
n
( x x ) 2 f
fL fR
i 1
i
0i
xi
xi x0
f
fC (x) f (x 0 ) Fi ( xi x0i )
x
i 1
i x0
n
{
1
if x0i (f / xi ) 0
Fi
x0i / xi
otherwise
• All second derivatives of fC are non-negative
• Called convex linearization (CONLIN), Claude
Fleury
Problems approximations
1. Construct the linear, reciprocal, and
convex approximation at (1,1) to the
3
1
f
1
function
x x x
2. Plot and compare the function and the
two approximations.
3. Check on their properties of convexity
and conservativeness.
1
2
1