CONCEPT OF OPTIMIZATION CLASSICAL

OPTIMIZATION TECHNIQUES

Chapter 1

Dr. Dereje Shiferaw

What is Optimization?
 Optimization is the mathematical discipline which is
concerned with finding the maxima and minima of
functions, possibly subject to constraints.
 Optimize means make as perfect, effective or
functional as possible
 Used to determine best solution without actually
testing all possible cases
Need for optimization in power system
 Power system operation is required to be
 Secure
 Economical
 Reliable
 All operations have to operate at optimum point
 Power flow analysis
 Economic Dispatch
 Reactive power
 Load shedding
 Configuration of electrical distribution networks etc
General Statement of a Mathematical Programming Problem
Find x which minimize: f(x)
Subjectgto:( x)  0 i  0,1,2,...h
i
li ( x )  0 i  h  1, h  2,...m

f(x), gi(x) and li(x) are twice continuously differentiable real

valued functions.
gi(x) is known as inequality constraint
X can be a vector with several
variables

Minimization of f(x) is same as

maximization of –f(x)
Some terminologies
 Design vector
 Design constraints
 Constraint surface
 Objective function
 Mathematical programming
Design vector
Two types of variables exist Design space- space of
-Pre-assigned variables
the design vector
-Variables
whose value is
known before hand f ( x )  9.82x1x 2  2x1
Design vector  x1  x2
- x 
-Vector of decision variables x 2 
Design vector Design space
-Should be calculated using
techniques

x1
Design constraints
 Restrictions on  Behavioral constraints
variables  Power cannot be
Find x which minimize: f(x) negative
Subject to:
gi(x) < 0 for i = 1, 2, ..., h  Geometric constraints
li(x) = 0 for i = h+1, ..., m
 Constraints due to
Where:
gi(x) is inequality constrain geometry
li(x) is equality constraint
Constraint surface
 Set of values which satisfy
a single constraint
 Plot of gi(x)=0
 Four possible points
 Free and acceptable
 Free and unacceptable
 Bound acceptable
 Bound unacceptable
Example
 Draw constraint surface for problem of minimization
f ( x )  9.82 x1x 2  2 x1 Subject to 1

0.9

2  x1  14 0.8

0.7

0.2  x 2  0.8 0.6

x2
0.5

0.4
2500
 500  0 0.3

x1x 2
0.2

0.1
2 4 6 8 10 12 14
x1

2500  2 (8.5x105 )( x12  x 22 )

 0
x1x 2 8(250) 2
Objective function
 Example – economic
 The function which gives
dispatch problem
the relation between the
 Minimize operating cost
objective we want to
 Variables – power
achieve and the variables output of each generator
involved  Constraint- system load
 Single or multiple demands, generating
capacity of generators
Objective function
 Amount of coal required
 Example: A power
company operates two
thermal power plants A
and B using three
different grades of coal
C1, C2 and C3. The
minimum power to be
generated at plants A and  Write the objective
B is 30MWh and 80MWh function to min cost
respectively.
 Objective function
surface
 Locus of all points
satisfying f(x)=C for
some constant C
 1

0.9

0.8

0.7

Red lines are 0.6

x2
objective function 0.5

surfaces for C=50 0.4

and C=30 0.3

0.2

0.1
2 4 6 8 10 12 14
x1
Classification of optimization problems
 Constrained optimization
1)Based on
existence of  Formulation
constraints Find X which Min F(X)
subject to
gj(X)0
 Unconstrained optimization
 Formulation
Min F(X)
 Classification cont…
2)Based on nature of  Static
design variables
 Design variables are simple variables
 Dynamic
 Design variables are function of other
variables
 Classification cont…

3)Based on nature  Linear programming

of expression for
objective function
 Objective function and constraints are linear
and constraints  Approximation
 Optimal Power flow, steady state security etc
 Nonlinear programming
 If any one of the objective function or
constraints is nonlinear
 Power system problems are nonlinear
Classification cont…

4)Based on
expression of
 Geometric programming – objective
objective function function and/or constraint are
or constrains
expressed as power terms
 Quadratic programming
 Special case of NLP
 Objective function is quadratic form
Classical Optimization techniques
Used for continuous and differentiable
functions
Make use of differential calculus

 disadvantages

Practical problems have non differentiable

objective functions
Single variable optimization
 Local minima
 If for Local
maxima
small positive and
negative h
 Local maxima
 If for
small positive and
negative h Local
minima
Single variable cont…
Theorem 1

Theorem 2
 Example
f (x)  12x 5  45x 4  40x 3  5 400

350

Soln. 300

250

Find f’(x) and then equate it with zero.

200

150
The extreme points are x=0,1 and 2
100

50
X=0 is inflection point, x=2 is local 0
minima and x=1 is local maxima -50

-100
-1 -0.5 0 0.5 1 1.5 2 2.5 3
Multivariable optimization
 Without constraint and With constraint
 Has similar condition with single variable case
Theorem 3

Theorem 4
Example: find extreme points of
 Function of two variable  Findextreme points
 Check the Hessian
matrix by determining
 Soln. second derivatives and
 Evaluate first partial determinants
derivatives and
Multivariable optimization with equality
constraints
 Problem formulation
 Find which minimizes F(x) subject to the constraint

gi(x)=0 for i=1,2,3, … m where m n

 Solution can be obtained using
 Directsubstitution
 Constrained variation and Lagrangian method
Direct substitution
 Converts constrained  Technique
optimization to  Express the m constraint
unconstrained variables in terms of the
 Used for simpler remaining n-m variables
problems  Substitute into the
objective function
Direct substitution cont…
 Example – find the values of x1,x2 and x3 which
maximize
subject to the equality constrain

Soln. Re-write the constraint equation to eliminate any

one of the variables
x 2  1  x12  x 32  then f (x1 , x 2 , x 3 )  8x1x 3 1  x12  x 32 
constrained variation method
 Finds a closed form expression for the first order
differential of f at all points where the constraints
are satisfied
 Example: minimize

 Subject to
Constrained variation
 At a minimum  Rewriting the equation

 If we take small  Substituting

variations dx1 and dx2
 Necessary condition
 After Taylor series expansion
Constrained variation
 For general case

 Under the assumption that

Example minimize the following function subject to given
constraint

 Minimize

 Subject to

 Soln. The partial  Using the necessary condition

differentials are
Method of Lagrangian multipliers

 Problem formulation
s.t.
 Procedure:
A function L can be formed as

 Necessary conditions for extreme are given by

Lagrange Multiplier method cont…

 For a general case L

Lagrangian method
 Sufficient condition is

Has to be positive definite matrix

 Example: find maximum of f given by

 Subject to

 Soln. The Lagrangian is  Giving

Formulation of multivariable optimization
 When the constraints are inequality constraints, i.e.

 It can be transformed to equality by adding slack

variable
 Lagrangian method can be used
Kunh- Tucker conditions
 The necessary condition for the above problem is

 When there are both equality and inequality constraints, the KT

condition is given us
Kuhn-Tucker conditions cont…

 Example: For the following problem, write the KT

conditions

Subject to
Linear programming
 History  Problem statement
 George B. Dantzing
1947, simplex method  Subject to the constraint
 Kuhn and Tucker –
duality theory
 Charles and Cooper -
industrial application
Properties of LP
 The objective function is  Transformations
minimization type  If problem is maximization
use –f
 All constraints are linear  If there is negative
equality type variable, write it as
difference of two
 All decision variables
 If constraint is inequality,
are nonnegative add slack or surplus
variables
Simplex algorithm
 Objective of simplex  Algorithm
algorithm is to find vector X 1. Convert the system of
0 which minimizes f(X) equation to canonical form
2.Identify the basic solution
3. Test optimality and stop if
and satisfies equality optimum
constraints of the form 4. If not, select next pivotal
point and re-write equation
5. Go to step 2
 Simplex algorithm
 Example: Maximize
 Solution:
Subject to
Step 1. convert to canonical form

Use tabular method to proceed on the

algorithm
Basic variable means those variables having coefficient 1 in one of the equation and
zero in the rest of the equations
Various types of solutions
 Unbounded solution
 Ifall the coefficients of the entering variable are
negative
 Infinite solution
 Ifthe coefficient of the objective function is zero
at an optimal solution
Modifications to simplex method
 Two phase method  Revised simplex method
 When an initial feasible  Solve the dual of the
solution is not readily basic solution
available
 Phase I is for
rearranging the
equations
 Phase II is solving
Using MATLAB to solve LP
 Example:
Subject to

Soln.
1. Form the matrices containing coefficients of the objective function, constraints
equations and constants separately
f=[5 2];
A=[3 4;1 -1;-1 -4;-3 -1];
b=[24;3;-4;-3];
lb=zeros(2,1);