Upload
adityashetgaonkar
View
223
Download
0
Embed Size (px)
Citation preview
8/9/2019 2. Classical Optimization Technique
1/19
1
Classical Optimization Techniques
Prof. Keyur P Hirpara
Assistant Professor
Relative and Global Optimum
A function is said to have a relative or local minimum atx =
x* if f(x*) f(x+h) for all sufficiently small positive and
negative values of h, i.e. in the near vicinity of the point x.
Similarly, a point x* is called a relative or local maximum iff(x*) f(x+h)for all values of h sufficiently close to zero.
A function is said to have a global or absolute minimum atx
= x* if f(x*) f(x) for all x in the domain over which f(x) is
defined.
Similarly, a function is said to have a global or absolute
maximum at x = x* if f(x*)f(x) for allx in the domain over
whichf(x) is defined.
2
8/9/2019 2. Classical Optimization Technique
2/19
8/9/2019 2. Classical Optimization Technique
3/19
Functions of a single variable
Necessary condition :
For a single variable function f(x) defined for x [a,b]
which has a relative maximum at x = x*, x* [a,b] if the
derivative f(x) = d f(x)/dx exists as a finite number at x =
x* then f(x*) = 0.
We need to keep in mind that the above theorem
holds good for relative minimum as well.
The theorem only considers a domain where the
function is continuous.
5
Functions of a single variable
The theorem does not say whathappens if a minimum or maximumoccurs at a point x* where thederivative fails to exist.
The theorem does not say whathappens if a minimum or maximumoccurs at an endpoint of the intervalof definition of the function. In thisexists for positive values of h only orfor negative values of h only, andhence the derivative is not defined atthe endpoints.
The theorem does not say that thefunction necessarily will have a
minimum or maximum at every pointwhere the derivative is zero. For example (fig), However, this point isneither a minimum nor a maximum.
6
8/9/2019 2. Classical Optimization Technique
4/19
Stationary points
Figure showing the three types of stationary points
(a) minimum (b) maximum (c) inflection point
In general, a pointx* at whichf(x*) = 0 is called a stationary
point.
7
Functions of a single variable
Sufficient condition:
For the same function stated above letf (x*) = f (x*)
= . . . = f
(n-1)
(x
*
) = 0, but f
(n)
(x
*
)
0, then it can besaid thatf (x*) is
a minimum value off (x) iff (n)(x*) > 0 and n is even;
a maximum value off (x) iff (n)(x*) < 0 and n is even;
neither a maximum nor a minimum if n is odd.
8
8/9/2019 2. Classical Optimization Technique
5/19
Example
Determine the maximum and minimum values of the function
f (x) = 12x5 45x4 + 40x3 + 5
f (x) = 60(x4 3x3 + 2x2)
= 60 x2(x 1)(x 2),
f (x) = 0 at x = 0, x = 1, and x = 2.
The second derivative is
f (x) = 60(4x3 9x2 + 4x)
Atx = 1, f (x) = 60
Here value of function is negative, hence atx = 1 is a relativemaximum.
fmax = f (x = 1) = 12
9
Example
Atx = 2, f (x) = 240
Here value of function is positive, hencex = 2 is a relative maximum.
fmin = f (x = 2) = 11
Atx = 0, f (x) = 0 andHere value of function is zero, hence we must investigate the next
derivative:
f (x) = 60(12x2 18x + 4) = 240 at x = 0
Sincef (x) = 0 at x = 0,
Here n=3 (even)
x = 0 is neither a maximum nor a minimum, and it is an inflection
point.
10
8/9/2019 2. Classical Optimization Technique
6/19
Example
Find the optimum value of the function f(x) = x2+3x-5and
also state if the function attains a maximum or a minimum.
Solution:
f (x) = 2x+3 for maxima or minima
OR x* = -3/2
f(x) = 2
Which is positive hence the pointx*= -3/2 is a point of minima
and the function attains a minimum value of -29/4 at this point.
11
Multivariable with No constraint
Necessary conditions
Iff(X) has an extreme point
(maximum or minimum) at
X = X and if the first
partial derivatives of f(X)
exist atX, then
12
8/9/2019 2. Classical Optimization Technique
7/19
Multivariable with no constraint
Sufficient Condition:
A sufficient condition for a stationary point X to be an extreme point is
that the matrix of second partial derivatives (Hessian matrix) off(X)evaluated at X is
Positive definite when X is a relative minimum point
And negative definite when X is a relative maximum point.
Consider the following second order derivatives
13
Multivariable with no constraint
A square matrix is positive definite if all its Eigen values are
positive and it is negative definite if all its Eigen values are
negative. If some of the Eigen values are positive and some
negative then the matrix is neither positive definite or negativedefinite.
To calculate the Eigen values of a square matrix then the
following equation is solved. |A-I|=0
14
8/9/2019 2. Classical Optimization Technique
8/19
Sufficient condition
Test that can be used to find the positive definiteness of a
matrix A of order n involves evaluation of the determinants
15
Nature of the extreme points
Point
X=X*
Value
(a11)
Value
H(x)
Nature of J Nature of X F(x)
(X1,X2) + Ve + Ve Positive definite
(Convex)
Relative
Minimum
Find the function at
minimum value
(X1,X2) + Ve - Ve Indefinite Saddle point Find the value of
function at [X1,X2]
(X1,X2) - Ve - Ve Indefinite Saddle point Find the value of
function at [X1,X2]
(X1,X2) - Ve + Ve Negative definite
(Concave)
Relative
Maximum
Find the function at
Maximum value
17
8/9/2019 2. Classical Optimization Technique
9/19
Nature of the extreme points
In the case of a function of two variables, f (x, y), the Hessian matrix may
be neither positive nor negative definite at a point (x, y) at which
In such a case, the point (x, y) is called a saddle point.it corresponds to
a relative minimum or maximum off(x, y)with respect to one variable,
say, x (the other variable being fixed at y = y) and a relative maximum
or minimum off(x, y)with respect to the second variable y (the other
variable being fixed at x).
As an example, consider the functionf (x, y) = x2 y2. For this function
These first derivatives are zero at x = 0 a n d y = 0. The Hessian matrix of f
at (x, y) is given by
18
Nature of the extreme points
Since this matrix is neither positive definite nor negative definite, the point
(x = 0, y = 0) is a saddle point.
The function is shown graphically in Fig.
It can be seen that f (x, y) = f (x, 0) has a relative minimum and f (x, y) =
f (0, y) has a relative maximum at the saddle point (x, y).
Saddle points may exist for functions of more than two variables also.
19
8/9/2019 2. Classical Optimization Technique
10/19
Example
Figure shows two frictionless rigid bodies (carts) A and B connected
by three linear elastic springs having spring constants k1, k2, and k3.
The springs are at their natural positions when the applied forceP iszero. Find the displacements x1 and x2 under the forcePby using
the principle of minimum potential energy.
20
Example
The potential energy of the system is given by
potential energy (U) = strain energy of springs work done by external
forces
The necessary conditions for the minimum of U are
The values of x1 and x2 corresponding to the equilibrium state,
obtained by solving above equations
21
8/9/2019 2. Classical Optimization Technique
11/19
Example
The sufficiency conditions for the minimum at (x1* , x2*) can
also be verified by testing the positive definiteness of the
Hessian matrix of U. The Hessian matrix of U evaluated at(x1* , x2*) is
The determinants of the square sub matrices of J are
22
Example
Find the extreme points of the function
The necessary conditions for the existence of an extreme point are
These equations are satisfied at the points
To find the nature of these extreme points, we have to use the sufficiency
conditions. The second-order partial derivatives offare given by
23
8/9/2019 2. Classical Optimization Technique
12/19
Example
The Hessian matrix of f is given by
Nature of the extreme point are as given below
24
Multivariable with Equality constraint
Direct Substitution method
Example: Find the dimensions of a box of largest volume that can
be inscribed in a sphere of unit radius.
Let the origin of the Cartesian coordinate systemx1, x2, x3be at
the center of the sphere and the sides of the box be 2x1, 2x2,
and 2x3. The volume of the box is given by:
f(x1, x2, x3) = 8x1x2x3
25
8/9/2019 2. Classical Optimization Technique
13/19
Since the corners of the box lie on the surface of the sphere ofunit radius, x1, x2, and x3 have to satisfy the constraint
This problem has three design variables and one equalityconstraint. Hence the equality constraint can be used toeliminate any one of the design variables from the objectivefunction. If we choose to eliminate x3,
Thus the objective function becomes
which can be maximized as an unconstrained function in twovariables.
Example
26
The necessary conditions for the maximum offgive,
Example
27
8/9/2019 2. Classical Optimization Technique
14/19
To find whether the solution found corresponds to a maximum or a
minimum, we apply the sufficiency conditions tof(x1, x2)
Since,
Hence the point (x1* , x2* ) corresponds to the maximum off.
Example
28
Solution by Constrained Variation Method
Variations about A
In above figure PQ indicates the curve at each point of which constraint issatisfied. If A is taken as the base point (x1* ,x2*), the variations in x1 and x2leading to points B and C are called admissible variations. On the other hand,the variations in x1 and x2 representing point D are not admissible since
point D does not lie on the constraint curve,g(x1, x2) = 0.Necessary condition
29
8/9/2019 2. Classical Optimization Technique
15/19
Example
A beam of uniform rectangular cross section is to be cut from a log having a
circular cross section of diameter 2a. The beam has to be used as a
cantilever beam (the length is fixed) to carry a concentrated load at the freeend. Find the dimensions of the beam that correspond to the maximum
tensile (bending) stress carrying capacity.
30
we know that the tensile stress induced in a rectangular beam () at any
fiber located a distancey from the neutral axis is given by
whereMis the bending moment acting andIis the moment of inertia of the
cross section about the x axis.
If the width and depth of the rectangular beam shown in Fig. are 2x and 2y,
respectively, the maximum tensile stress induced is given by
Subject to the constraint
x2 + y2 = a2
This problem has two variables and one constraint;f = kx1y2
g = x2 + y2 a2 (Where, k= 3M/4)
Example
31
8/9/2019 2. Classical Optimization Technique
16/19
We have equation,
Put this value in constraint,
Example
32
Solution bythe Lagrange Multipliers method
Take example of two variables and one constraint
Minimize f (x1, x2)
Subject to, g(x1, x2) = 0
The necessary condition for the existence of an extreme point atX =X* was found in previous Section
(Eq-1)
By defining a quantity , called theLagrange multiplier, as
(Eq-2)
Equation can be expressed as
(Eq-3)
33
8/9/2019 2. Classical Optimization Technique
17/19
Solution bythe Lagrange Multipliers method
Equation 2 can be expressed as
(Eq-4)
Hence equations (3) to (4) represent the necessary conditions
for the point [x1*, x2*] to be an extreme point.
Necessary conditions require that at least one of the partial
derivatives ofg(x1, x2) be non-zero at an extreme point.
The necessary conditions given by Eqs. (3) to (4) are more
commonly generated by constructing a function L, known asthe Lagrange function, as
34
Solution bythe Lagrange Multipliers method
By treating L as a function of the three variables x1, x2, and ,
the necessary conditions for its extreme are given by
35
8/9/2019 2. Classical Optimization Technique
18/19
Find the solution of Previous example using the Lagrange multiplier
method:
The necessary conditions for the minimum off (x, y)
Comparing both values of derived from first two equations and put in to
third equation,
Example
36
Necessary Condition for a GeneralProblem
The Lagrange function, L, in this case is defined by
introducing one Lagrange multiplier j for each constraint gj(X) as
37
8/9/2019 2. Classical Optimization Technique
19/19
Sufficient Condition for a GeneralProblem
Where
If determinant of this equation is positive function is Minimum.
If determinant of this equation is Negative function is Maximum.
If some of the roots of this polynomial are positive while the others arenegative, the point X* is not an extreme point.
38
Multivariable optimization withinequality constraint