B.M.S.

COLLEGE OF ENGINEERING, BENGALURU – 560 0 19

Autonomous college, affiliated to VTU
DEPARTMENT OF MATHEMATICS

Unit-5
Multivariable Optimization
Constrained Optimization
Lagrange Multipliers:
Property: The gradient of a function is perpendicular to the contour lines, the
tangents to the contour lines of and are parallel if and only if the gradients
of and are parallel.
Thus we want points (x, y) where g(x, y) = c and f   igi  0

Method of Lagrange multipliers (Equality constraints)

Given,  subject to 
Step–1: Write the Auxiliary Equation 
Step–2: Find , and by differentiating partially w.r.t. ‘ and respectively.

Step–3: Solve the equations and together with

Step–4: The values of and so obtained will give the stationary value of
Note: For multiple constraints

 Subject to then

Auxiliary Equation 
Follow the same steps as we defined above

Problems:
1. Find the maximum and minimum distances of the points (3, 4, 12) from the sphere
x2  y 2  z 2  1 .
2. Find the maximum and minimum of f ( x, y )  5 x  3 y subjects to the constraints
x 2  y 2  136 .
3. Find the maximum and minimum values of f ( x, y, z )  3x 2  y subjects to the constraints
4 x  3 y  9 and x2  z 2  9 .
4. Find the maximum and minimum values of f ( x, y, z )  4 y  2 z subjects to the constraints
2 x  y  z  2 and x 2  y 2  1 .
 5. Find the maximum and minimum values of f ( x, y, z )  2 x  4 y  z 2 subjects to the
constraints y 2  z 2  1 and x 2  z 2  1 .
6. Find the maximum and minimum values of f ( x, y, z )  x  y  z 2 subjects to the
constraints x  y  z  1 and x 2  z 2  1 .
7. Find the extreme values of f ( x, y, z )  2 x  3 y  z subjects to the constraints
x 2  y 2  5 and x  z  1 .
8. An asteroid is entering the atmosphere of moon. The shape of the asteroid is described
by the equation 4x2 +y2 +4z2 = 16. The temperature on the surface of the asteroid after
one month was observed to be represented by equation 8x2 + 4yz − 16z + 600. Is it
possible to find the point on surface of asteroid with maximum temperature? If yes, find
it?
Karesh-Kuhn-Tucker (KKT) optimality conditions:
(Inequality constraints)
For the problem:

Min subject to variables, m constraints)

The necessary conditions are:

f   igi  0 (optimality)

for i = 1, 2, ..., m (feasibility)

for i = 1, 2, ..., m(complementary slackness condition)

for i = 1, 2, ..., m (non-negativity)

Note: Here first condition is called Stationary condition, second is complimentary

slackness, third is Feasibility condition and last one is non-negativity condition.
Problems:

1. Apply KKT method to solve NLPP maximize

subject to the constraint

2. Apply KKT method to solve NLPP maximize subject to the

constraint

3. Apply KKT method to maximize Z  12 x1  21x2  2 x1 x2  2 x12  2 x22 subject to the

constraints x1  x2 10; x2 8, x1 , x2 0.

4. Apply KKT method to solve NLPP maximize subject to the

constraint

5. Maximize f ( x, y )  xy subject to the constraint x  y 2 2; x, y 0 .

Unconstrained Optimization
Fibonacci Method
Fibonacci numbers
The Fibonacci search is based on the sequence of Fibonacci numbers which are defined by the
equations

Thus the Fibonacci numbers are 1, 1, 2, 3, 5, 8, 13, 21, 34 · · ·.

Algorithm
Step1: Choose Lower bound ‘a’ upper bound ‘b’
 Set

Step 2: Compute ( )

Set
Step 3: Compute
Step 4: > If then take new interval as [ ] [ ]
> If then take new interval as [ ] [ ]
If then set go to step 2; else terminate.
Examples:

1. Apply Fibonacci method to minimize the function over [ ] by taking

2. Apply Fibonacci method to minimize the function in the range by

taking

3. f ( x)  x4  14 x3  60 x 2  70 x Use Fibonacci search method to find the value of x that

minimizes f over the range [0,2].

4. Apply Fibonacci method to minimize the function in [0,1] within the

interval of uncertainty of the initial interval of uncertainty.

5. Minimize [ ] using Fibonacci search perform 6

iterations.

1  x 1 
6. Use Fibonacci algorithm to minimize f ( x)  2 
log x  2 .
( x  1)  x 1 

Three-point interval search method

Algorithm
Step 1 Objective function is to be maximized

Step 2 Insert three equally spaced points in [a,b] , these points can be calculated as

i.e

Step 3 calculate

Step 4 Identify maximum values of

 If maximum value is at then we can’t solve it further.

  If maximum value is at interior points say then function is to be optimized in
[ ]
 Repeat the steps 2,3 and 4 until the solution
Problems

1. Apply three-point interval search method to find Maximize on [ ] where

2. Apply three-point interval search method to find Maximize on [ ]

where

3. Apply three-point interval search method to find Maximize on [ ]

where

4. Apply three-point interval search method to find Maximize on

[ ] where