Problem 6 Problem Set 2.3A Page 26(Modified)

Electra produces two types of electric motors,

each on a separate assembly line. The respective

daily capacities of the two lines are 150 and 200

motors. Type I motor uses 2 units of a certain

electronic component, and type II motor uses

only 1 unit. The supplier of the component can

provide 400 pieces a day.The profits per motor

of types I and II are $8 and $5 respectively.

Formulate the problem as a LPP and find the

optimal daily production.

Let the company produce x1 type I motors

and x2 type II motors per day.

1 28 5z x x

Subject to the constraints 1

2

1 2

1 2

150

200

2 400

, 0

x

x

x x

x x

The objective is to find x1 and x2 so as to

Maximize the profit

We shall use graphical method to solve the

above problem.

Step 1 Determination of the Feasible solution space

The non-negativity restrictions tell that the

solution space is in the first quadrant.

Then we replace each inequality constraint

by an equality and then graph the resulting

line (noting that two points will determine a

line uniquely).

Next we note that each constraint line

divides the plane into two half-spaces and

that on one half-space the constraint will be

≤ and on the other it will be ≥. To determine

the “correct” side we choose a reference

point and see on which side it lies.

(Normally (0,0) is chosen. If the constraint

line passes through (0,0), then we choose

some other point.) Doing like this we would

get the feasible solution space.

Step 2 Determination of the optimal solution

The determination of the optimal solution

requires the direction in which the objective

function will increase (decrease) in the case of

a maximization (minimization) problem. We

find this by assigning two increasing

(decreasing) values for z and then drawing the

graphs of the objective function for these two

values. The optimum solution occurs at a point

beyond which any further increase (decrease)

of z will make us leave the feasible space.

Graphical solution

z=1800

z=1700

z=1200

z=400

z=1000

(100,200)

(150,100)

(150,0)

(0,200)

x1

x2

Maximize z=8x1+5x2

Subject to the constraints

2x1+x2 400

x1 150

x2 200

x1,x2 0

Optimum

=1800 at

Feed Mix Problem

Minimize z = 2x1 + 3x2

Subject to x1 + 3 x2 15

2 x1 + 2 x2 20

3 x1 + 2 x2 24

x1, x2 0

Optimum = 22.5 at (7.5, 2.5)

(7.5,2.5)

(4,6)

(0,12)

(15,0) Minimum at

z = 42

z = 39

z= 36

z = 30 z = 26

z = 22.5

Graphical Solution of Feed Mix Problem

Maximize z = 2x1 + x2

Subject to x1 + x2 ≤ 40

4 x1 + x2 ≤ 100

x1, x2 ≥ 0

Optimum = 60 at (20, 20)

(20,20)

(0,40)

(25,0)

z maximum at

Maximize z = 2x1 + x2

Subject to 2

1 2

1 2

1 2

1 2

10

2 5 60

18

3 44

, 0

x

x x

x x

x x

x x

[1]

[2]

(5,10)

[3]

(10,8)

[4]

(13,5)

(14.6,0)

(0,10)

z is maximum at (13, 5)

Max z = 31

z = 4

z = 10

z = 20

z = 28

z = 31

Exceptional Cases

Usually a LPP will have a unique optimal

solution. But there are problems where

there may be no solution, may have

alternative optimum solutions and

“unbounded” solutions. We graphically

explain these cases in the following slides.

We note that the (unique) optimum

solution occurs at one of the “corners” of

the set of all feasible points.

Alternative optimal solutions

1 210 5z x x

Subject to the constraints 1

2

1 2

1 2

150

200

2 400

, 0

x

x

x x

x x

Consider the LPP

Maximize

Graphical solution

z=2000

z=1500

z=400

z=1000

(100,200)

(150,100)

(150,0)

(0,200)

x1

x2

Maximize z=10x1+5x2

Subject to the constraints

2x1+x2 400

x1 150

x2 200

x1,x2 0

z maximum

=2000 at

z=600

Thus we see that the objective function z

is maximum at the corner (150,200) and

also has an alternative optimum solution at

the corner (100,200). It may also be noted

that z is maximum at each point of the line

segment joining them. Thus the problem

has an infinite number of (finite) optimum

solutions. This happens when the objective

function is “parallel” to one of the

constraint equations.

Maximize z = 5x1 + 7 x2

Subject to

2 x1 - x2 ≤ -1

- x1 + 2 x2 ≤ -1

x1, x2 ≥ 0

No feasible solution

Maximize z = x1 + x2

Subject to

- x1 + 3 x2 ≥ 30

- 3 x1 + x2 ≤ 30

x1, x2 ≥ 0

unbounded solution z=20

z=30

z=50

z=70

Let x = (x1, x2) and y = (y1, y2) be two pints in

the x1 x2 plane. Any point t = (t1, t2) on the line

segment joining them is of the form

t1 = (1 - ) x1 + y1

t2 = (1 - ) x2 + y2

for some : 0 1

= 0 corresponds the „left‟ endpoint x

= 1 corresponds the „right‟ endpoint y

More generally, let x=(x1, x2, …, xn) and

y=(y1, y2, …, yn) be two points in the n-

dimensional space Vn. The line segment

joining x and y is the set of all points of

the form t = (1 - ) x + y, for all :

0 1 in the sense that the ith

coordinate of t, namely ti is given by

ti = (1 - ) xi + yi for all i = 1, 2, …n.

Convex sets

A subset S in the n-space Vn is called

convex if x, y belong to S implies the line

segment joining them also lies in S.

convex subset NOT convex

In Vn, the “half spaces”

x=(x1, x2, …, xn) | a1x1+a2x2 + …+ anxn ≤ b

x=(x1, x2, …, xn) | a1x1+a2x2 + …+ anxn ≥ b

x=(x1, x2, …, xn) | a1x1+a2x2 + …+ anxn = b

are all convex.

and the “hyperplane”

Theorem: The intersection of any number

of convex sets is a convex set.

Corollary: The set of all feasible solutions

of an LPP is a convex subset.

Extreme Points (Vertices=corners)

Let S be a convex set. A point t in S is

called an extreme point of S if it is not

strictly between two distinct points of S.

In other words, whenever x, y are two

points in S, we cannot find a : 0 < < 1

such that t = (1 - ) x + y

Extreme point Every point on the

boundary is an Extreme

point

Extreme point

Existence of extreme points

Theorem: Let S be a nonempty, closed,

convex set that is either bounded from

above or bounded from below. Then S

has at least one extreme point.

Extreme Points and LPP

Suppose the set SF of all feasible solutions

of a LPP is nonempty and bounded. (We

already “know” it is closed and convex.)

Then the optimum solution of the LPP

occurs at a corner (=extreme point) of SF.

Proof: You may see the book by JC Pant.