0% found this document useful (0 votes)
17 views20 pages

Optimization, Chapter 2

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
17 views20 pages

Optimization, Chapter 2

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 20

3.2 Problem 3.

3
Optimization Techniques Linear Programming and Tran sportation

he word
'programming' does not refer here to computer a X,t a, X,
t..
ta X =b,
ynonym tor
planning.Thus mathematical programming-rather
program is a planning of activities. In
a it is
..ta2 X b,
mathematical program, we seek any 21X a22 X
t

a solution of
planning. (2)
Ihus, linear
programming is an optimization method applicable for the solution
problems in which the
the decision objective function and the constraints appear as linear of
equations in a LPP may be in the form functions of
variables. The constraint
or inequalities.
of equalities
3X, am2 X t . tamn X, =bm
George B.
Dantzig, formulated the general LPP and devised the X a . . X, 20
(3)
and
solution in 1947.
Although several other methods have been simplex method of and the decision
solving LP problems, the simplex method developed over the years for and a,(1= 1 to m, j
=
1 to n) are known constants, and x,
continues to be the most efficient and
where c b,
method for solving popular variables.
general LP problems. Linear
that permits us to make optimal decisions in programming is a revolutionary development 2. Matrix form
complex situations. (1)
Some areas of application of LPP: The number of Optimize
fX) = cTX

large that it is not possible to describe all ofapplications


has been so of linear programming8
them here. Only some subject to the constraints
are mentioned here: applications aX = b (2)
Oil and gas
industry: The initial extraction of the crude oil and gas from the earth, and X 20 (3)
finding of the operating schedules of a refinery, obtaining the
when blending gasolines, the last optimal product mix
phase of transporting the refined product to the
consumer points.
where
ii. Optimization of the
mixing of livestock feed
i. The shipment of finished
goods from a number of plants to a number of ware houses
iv. The using of forest goods a1 a12 ain
effectively
a21 a22 a2n
V. Scheduling of production in many manufacturing units or industries
vi. The allocation of pilots and crew to air crafts
dml am2 ämn
vii. Scheduling of flight times of airlines
Thus, the reformulated general be put in the following matrix form:
LPP can
vii. The distribution of resources for advertising
Find X such that, it optimizes the linear function f(X)
cTX subject to the constraint
=

ix. The portfolio selection of shares and stocks


aX = b and X 2 0.
x. The assignment of jobs to people, and many other such problems
constraints aX b is known a
Remark: The form, Maximize f(X) cTX, subject to the
=
=

pefinition: The general LPP calls for optimizing (maximizing/minimizing) a linear function the standard form of LPP. The form with constraints aX bis called its canonical form
of variables called the 'objective function subject to a set of linear equations and/or
the constraints restrictions. Characteristics of the standard form of LPP
inequalities called or

Now, consider: The characteristics of a maximization LPP are

i. The objective function is of the maximization type


3.1.2 Standard form of linear Programming problem restrictior

be stated in the standard form: l All the constraints are of the equality type except for the non-negativity
The general LPP can following which remain as in equalities
1. Scalar form I. The right side element of each constraint is non-negative
Optimize f(x, X ..., X,) =C, x, t c, X, t... tC, X, (1) IV All the decision variables are non negative.
subject to constraints
Problem
3.5
3.4 Transportation
Optimization Techniques Linear Programming and

Before going to solve the given LPP, it should be


expressed the standard
using following transformations: in
form by Solved prob
Problem 1:
1The minimization form
of a function f(x,, Xq, .-.,
x) is equivalent to the maximization of the Express te
negative of the same function.
Minimize f-x,-2x, +*,
For example, the objective function subject to 2x,+ 3x,+ 4x,2-4
Maximize f=c, x, + c, x, t ...+c, x, 3x,+Sx, +2x, 27
is equivalent to X, 20, X,20, X,
is in sign
unrestricted
and the standard LPP 1s:
above transformation rules,
minimize
f-f-c, Proceeding according to the
x, -C, X, C , X Selution:
()-f-x, +2x, -(x,-x,")
Consequently, the objective function can be stated in the maximization form Maximise
-2x, -3x,-4 (x,- x,") s, = 4
in any +
LPP. subyect to
ii. Constraints of the inequality (s or 2) type can be 3x,+5x, +20x,- x,") -s, =7
changed to cquations ( ) by
augmenting (adding or subtracting) the left side of each such constraint by non-negative and XX X", x, S, 5, 20
ariables. increased to six.
clearly, the number of variables
If a constraint appears in the form of 's' type of 3.1.3 Formulation of LP problems
a
inequality as
collection ofrelevant
,,X,ta X, t... a,. X, s b, n general, in optimization theory,
after identi fication of the problem,
translated into appropriate mathematical mouc
data, the given problem should be mathematical model of LPP
includes
vertcd into the equality form by adding a non neeative slack varisht.
callcd formulation. The
'S, as follows: process of translation is

three basic elements:


X,t X, t . + a,, x, + s, b , 1Deciston variables that we seek to determine
Similarly, if the constraint is in the form of a "2' type of inequality as to optimiZe
Objective function (goal) that
we aim

, X,+ X, tta,, X, 2 b, Constraints (restrictions) that we need to satisfy


toward the
variables is an essential first step
it can be converted into equality form by subtracting a non negative surplus variable The proper definition of the decision
variables defined, the task of constructing the
S, as the are
development of the model. Once
should not be too difficult.
objective function and the constraints
X,ta X, t t a , X, -5, =b for mathematical formulation of a
LPP consists of the following
Thus, the procedure
ii. The right side can always be made positive on multiplying both sides of the resulting
equation by -1 whenever it is necessary.
major steps:
find the key decisions to be made.
Siep 1: Study the given LPP to
In most engineering problems, the decision variables x, will be nonnegative. determined and designate them by algebraic
iv. But, a Step 2: Identify the decision variables to be
variable may be unrestricted in sign in some problems. In such cases, an unrestricted
the symbols Ilinear inequalities or
variable (which can take a positive, negative, or zero value) can be written as
constraints in the problem and expres
them as

difference of two nonnegative variables. Step 3: identify the


cquations linear function of the decision
Thus if x, unrestricted in sign, it be written as function and express it as a
is can
Step 4: Identify the objective according as it is a cost function
minimized or maximized
This function has to be
Xx" arta bies.
or profit function
where x 0 and x" 20
variables
Step 5: State the conditions of En
It can be seen that x, will be negative, zero or positive, depend1ng on whether x is 2
situations are concerned with the problem of planning activity.
Most of the practical is to make such a
at our disposal and our problem
=
or S x cach case, there are limited resources

production or to
the cost of
minimize

use of these resources as to yield the maximum


Problem 3.7
3.6 Linear Programming
and Transportation
Optimization Techniques

to maximize
production, or to give the maximum profits, ete. In this section, illustration of the applicatio Find x,, X, and x, $o as
of LP to various situations will be shown. In these examples the stress is laid on f 2x, + 4x, + 3%,
th
analysis of the situation and formulation of the linear programming problem rather tha subject to the constraints
on its solution. s 3600
3x, + 4x,+ 3x,
2400
Solved problems 2 2x,tx,t 3x, s
Problem 1: x+3x,+ 3x, s 4800
20
A company has three operational departments (weaving, processing and packing) wit and x, 20, X, 20, x,
capacity to produce three different types of clothes namely suitings, shirtings and wollen Problem 2: requirements
yielding a profit of Rs. 2, Rs. 4 and Rs. 3 per meter respectively. One meter suiting a diet which
will fulfill his daily
decide the constituents of
wants to is to be made from
requires 4 minutes in weaving, 3 minutes in each department. In a week, to tal run time of A person cost. The choice
carbohydrates at the minimum
of proteins fats and foods are given in
the following
of these
each department is 60, 40 and 80 hours for weaving, processing and packing department foods. The yields per
unit
four different types of
respectively.
table:
Formulate the linear programming problem to find the product mix to maximize thd Yield/Unit
Cost/unit
profit.
Solution: The data of the problem is summarized below: Food Rs.
Fats Carbohydrates
Type Proteins
Profit 45
Departments
WeavingProcessing Packing (Rs. per meter) 40
4
in min.) (in min.) (in min.) 85
8
65
Suitings 4 6 5
Shirtings 4 4
Minimum 800 200 700

Wollens 3 3 3
Requirement
availability (min) 60 x 60 40x60 80x 60
the problem.
Formulate the LP model for
Step 1: The key decision is to determine the weekly rate of production for the three
types of clothes. Solution: minimum cost.
is to select a food type which is having
Step 2: Let us designate the weekly production of suitings, shirtings and wollens by x,, X, Step 1: The key decision respectively
the four food types as x, X X, and x.
and x, meters respectively Step 2: Designate 1,2,3 and 4 the person requir
constraints: From these units of food type
Step 3: The constraints are the limited availability of three operational departments. One Step 3: Going to frame
3x, + 4x, +8x, 6x,proteins/day
+
meter of suiting requires 3 minutes of weaving. The quantity being x, meters, the requirement
for suiting alone will be 3x, units. Similarly, x, meters of shirting and x, meters of woolen + 5x, fats/day
2x, +2x,+7x,
will require 4x, and 3x, minutes respectively. Thus the total requirement of weaving will be carbohydrates/day
6x, + 4x, + 7x, 4x,
+
800, 20
3x, +4x, + 3x,, which should not exceed the available 3600 minutes. So the labour constraint fats and carbohydrates are
becomes 3x, + 4x, + 3x, s 3600 similarly, the constraints for the processing department and Since the minimumrequirement of these proteins,
the constraints are
packing departments are 2x, + x, + 3x, s 2400 and x, + 3x, + 3x, 8400 respectively. and 700 respectively,
+ 6x,2 800
3x, 4x, +8x, t
Step 4: The objective is to maximize the total profit from sales. Assuming that whatever is 200
2x, +2x, +7x, Sx, 2
produced is sold in the market, the total profit is given by the linear relation f= 2x, + 4x, + 3x, t

2 700
Step 5: Since it is not possible to produce negative quantities, then x,, X,, X,20. 6x, + 4x, 7x, 4x,
+ t

The LPP can thus be put in the following algebraic format: and , Xg X, and x, 2 0
Problem
3.9
tion
and Transporta
Linear Programming
3.8 Optimization Techniques
Magazine
Step 4: Since, the costs of these food of type 1,2, 3 and 4 B C
Rs. 45, Rs. 85 and Rs. 65
are A
per unit, the total cost is Rs.
total cost, the
45x, + 40x, + 85x, + 65x,. As the objective is to minimize the 0.6 Lakh 0.4 Lakh
objective function is 1 Lakh
Readers 8%
Minimize f= 20% 15%
45x, +40x, + 85x, + 65x Principal Buyers 5000
Step 5: Cost per Advt (Rs.) 8000 6000
XX Xg and x, 2 0 has alread
Now, the LPP is Find x, X,, X, and x, so as to advertisements. The owner
Rs. 1 lakh for the
is atmost and
advertisements and that B
The budgeted amount
Minimize f=45x, + 40x, t 85x, + 65x, A should have no more than 15
decided that magazine. the problem.
on LP model for
subject to 3x, + 4x,+8x, t 6x, 2 800
cach have at least 80
advertisements Formulate

2x, + 2x, + 7x, + 5x, 2 200


Solution
the total exposure (due to advt.)
6x,+ 4x + 7x, + 4x,2 700 decision: Maximize
Key A
and 2 0 = number of insertions in magazine
X X2 X, X Decision variables: x
B
Problem 3: -do-
The manager of an oil -do-
refinery
must decide on the optimum mix of two
possible blending
processes of which the input and output production runs are as follows: l00000 (Budget)
Ccnstraints: 8000 x, + 6000x, + 5000x, <
8 (Ads)
Input Output x, S15, x,28, x, 2
Process x, 20, x,20 and x,20
Crude A Crude B Gasoline X Gasoline Y
Objective function
6 9
Maximize (total exposure)
5 (20% of 10,0000)x, +(15%
= of 60,000)x, +(8% of 40,000)x,
The maximum amounts available of crudes A and B are 250 units and 200 units = 20,000x, +9000x,+ 3200x,

respectively. Market demand shows that at least 150 units of gasoline X and 130 units of Problem
all tomatoes, lettuce,
or radishes he can rai
gasoline Y must be produced. The profits per production run from process 1 and process A farmer has 100 acre farm. He can sell fettuce a
for tomatoes, Rs. 0.75 a head for
2 Rs. 4 and Rs. 5 The price he can obtain is
Rs. 100 per kg
are respectively. Formulate the problem for maximizing the profit. acre is 2000 kg
of tomatoes, 30
The average yield per
Solution: Key decision: maximizing the profit, RS. 2.00 per kg for radishes. Rs. 0.50 per kga
Fertilizer is available at
heads of lettuce and 1000 kgs of radishes. and 50 kgs
Decision variables: x, = number of units of gasoline from process 1
each for tomatoes and lettuce,
acre is 100 kgs
the amount required per acre is S
man-d:
and harvesting per
x,number of units of gasoline from process 2 Labour required for sowing,
cultivating
man-days of lab
Tadishes. for lettuce. A total of 400
6x, + 5x, S 250, 200 and 6 man-days
constraints :
4x, +6x, S Tor tomatoes and radishes,
of labour are available
at Rs. 20 per
man-d

6x, +5x, 2 150, 130 are available at Rs. 20 per man-days


9x, + 5x,2 LPP to maximize the tarmer's total protit.
Formulate this problem as a
and K20
Solution: should be allotted
to e
objective function : Maximize f= 4x, t
5x is to decide how much
area

ey decis1on: Farmer's problem


Problem 4: to m a x i m i z e his total protit.
of rop he wants to grow
ype
The owner of cosmo sports whishes to determine how many advertisements to in
place Decision variables
the selected three monthly magazines A, B and C. His objective is to a advertise in such Number of acres to grow tomatoes

way that total exposure to principal buyers of expensive sports good is maximized. X
-do- lettuce
Percentages of readers for cach magazine are known, Exposure in any particular magazine radishes
-do
is thenumber of advertisements placed multiplied by the number of principal buyers. The
following date may be used
3.10 Problem 3.11
Optimization Techniques Linear Programming and Transporta tion

Objective Function: Decision variables:

x,= Number of Grade inspectors


As the farmer will produce 2000x, kgs of tomatoes, 3000x, heads of lettuce, and 1000x, I
kgs of radishes Number of Grade II inspectors
x,=
Total sale R s (2000x, +0.75 3000x, + 2
x x
1000x,) Constraints:
we have the following
Fertilizer expenditure Rs (0.50[100(%, + x,) + 50x,) Since the number of available inspectors in each grade is limited,
Labour expenditure =
Rs[20 (5x, +6x,+ 5x,)]
x
constraints

Therefore, farmer's net profit will be X, S8 (Grade I)


P= Total sale (in Rs) -Total expenditure (in Rs.) xS 10(Grade II)
thus get
least 1800 pieces to be inspected daily,
we
P [2000x, +0.75 x 3000x, + 2x 1000 x, The company requires at
-

0.50 x
[100 (x, x,)+ 50x,]
+ -

20 x
[5x, +
6x, +
Sx,] 8(25)x,+ 8(15)x, 2 1800
=

1850x, +2080x, + 1875x, 5x,+ 3x, 45


Constraints : Objective function:
that the company incurs two
types of costs
Total area of the farm is restricted to 100 acre To develop the objective function, we note errors. The
paid to the inspector, and the cost of his inspection
X,x+ x, S 100 during inspection: wages
of each Grade I inspector 1s
total man-days labour is restricted to 400 man-days hour!y cost
Rs. [4+ 2(25) (0.02)] = Rs. 5 per hour
Sx, +
6x, +5x, s 400
similarly, for each Grade II inspector
and
Rs[3+2(15) (0.05)] =Rs. 4.50 per hour
Hence the farmer's allocation problem can be finally put in the form:
Thus the objective function is to minimize the daily cost of inspection given by
Find the value of x, X2 X, S0 as to maximize
f= 8(5x,+ 4.50x.) = 40x, + 36x,
P 1850x, + 2080x, + 1875x,
subject to the conditions
The complete formulation of the LPP thus becomes
Minimize f= 40x, + 36x,
100
X,+, t X, S subject to X, S 8
Sx, +6x, + 5x, < 400
x,s 10
and

Problem 6: 5x, +3x, 45


A company has and X X,0
two grades of inspectors 1 and 2, who are to be assigned for a quality
Note: Mathematical formulation (modeling) is not a science but primarily an art and
control inspection. It is required that at least 1800 pieces be inspected per 8-hour day
Grade I inspectors can check pieces at the rate of 25 per hour, with an accuracy of 98%.
comes mainly by practice.
Grade II inspectors check at the rate of 15 pieces
per hour, with an accuracy of 95% 3.1.4 Geometry of linear programming problemss
The wage rate of Grade I inspector is Rs. 4.00 per hour, while that of Grade II inspector case, and it can be
Iinear programming problem with only two variables is a simple
is Rs. 3.00 per hour. Each time an error is made by an inspector, the cost to the company method
gives
a

sOIved by using graphical method. Apart from the solution, the graphical
is Rs. 2.00. The company wants to determine the optimal assignment of inspectors, which linear programming probiems
Pysical picture of certain geometrical characteristics of Some ot the
will minimize the total cost of the inspection. nc geometrical characteristics can also be studied mathematically.
of LPP
Solution: powerful methods of LPP take advantage of these characteristics. beroc
ore solving we have to
two variables,
Key decision: Determination of the optimal assignment of inspectors, which will minimize consider graphical method to solve LPP with only
the total cost of the inspection a eto
8g r the terminology in linear programming and some of the imnportant theorciud
Problem 3.15
3.14 Optimization Techniques Linoar Programming and Transportation

LP problem, if it satisfies
Solution to LPP: Any set X of variables is calleda solution to
the set of constraints (2) only
Thus, x eand X e
called a feasible solution of LP problem,
Feasible solution (FS): Any set X of variables is
conditions (3) also.
if it satisfies the set of constraints (2) and non negativity
AX)+ (1 - ) X ellc Basic solution: A basic solution to the set of constraints (2) is a
solution obtained by
to zero and solving the constraints
setting any n- m variables (among n variables) equal
variables (of course, some of them may be zero) are

(2) for remaing m variables, such m


Hence by definition of convex setc, is a convex set Hence the proof. valued variables are called non-basic
called basic variables and remaining n-m zero

Note : If there are a number of convex sets represented by c,, c, ... the set of points c variables.
n!
common to all these sets will also be convex.
will be at the most n m)! m!
The number of basic solutions thus obtained (n -

things taken at a time.


which is the number of combinations of n
m

to zero to obtain the basic solution is


Basis: The collection of variables not set equal
called the basis.
which also satisfies the non-
Basic Feasible solution (BFS): A BFS is a basic solution
Intersection of two convex sets negativity conditions (3), that is all basic variables are non-negative.
Basic feasible solutions are of two types:

Theorem : The feasible region of a linear programming problem is convex degenerate BFS is the BFS
which has exactly m
Non degenerate BFS: A non

Proof: The feasible region F of a standard LPP is defined as positive x, basic variables are
if one or more
F {X| aX = b, X2 0} (1) 11. Degenerate BFS: A BFS is called degenerate,
zero-valued
Let X, XEF, then if it also optimizes objective the function (1)
Optimum BFS: A BFS is said to be optimum,
aX =b, X20 (2)
function (1) can be increased or
Unbounded solution: If the value of the objective
aX =b, X20 (3)
decreased indefinitely such solutions are called
unbounded solutions.
consider x (2) + (1 - ) * (3):
feasible solution. However, an optimum
Note: Unless otherwise stated, solution means a minimum.
finite maximum or finite
a[X + (1 -

2) X®] =ab + (1 - 2)b =


b solution to a LPP imply that objective function has a

that is aX = b where X) = AX) + (1 -1)X©)


Solved Problems 3
Thus the point X satisfies the constraints and if 0 < 1 s 1, X0 2 0
Problem 1:
This means that XO)e F and consequently F is convex. Is the non-

Obtain all the basic solutions to the following system of linear equations?
Hence the proof. degenerate solution feasible?
Theorem: Any local minimum solution is global for a LPP. 2x +6x,+ 2x, t x,=3
6x, t 4x, + 4x, + 6x,=2
Consider the matrix form of LPP:
Is the ne
optimize f= cTX (objective function) (1) which of them are BFS and which are non degenerate basic solutions'?
degenerate solution feasible?
subject to aX b, b 20 (constraints) (2)
Solution: here n= 4, m 2, n- m= 2 and 4= 6
and X20 (non negativity conditions) (3) 6 different basic solutions are given in the following table:
3.17
Problem
3.16 Optimization Techniques and Transporta tion
Linear Programming

Basic Non Basic Values of feasible? non degenerate? feasible &Problem 3:


s i m u l t a n e o u s equations
SNo variables| variables Basic non solutions for the system of
Find all basic
variabless degenerate 2x, +3x,-2x, - 7x,=1
x,X0x,0
t =4
XX, X, 5x, number of basic
Yes No No maximum possible
3, n= m
=
l and then the
=4, m
-

Solution: here n we will have


variable equal to zero, each time,
solutions is 4c 4, by taking one
No Yes No = 0

X, 0

XXX, =X, =0 X
F 0

X No Yes No x-0X,= solutions


set of feasible
Theorem: Every BFS is an extreme point of the convex
linear function
Then the minimum of a
X, X X, X, = 0X 4 Theorem: Let C be a closed
c o n v e x polyhedron.

extreme point of C.
x, 0 Yes No No over C is attained at an
VARIABLES
GRAPHICAL PROCEDURE TO
SOLVE LPP OF TWO
method. The method
X Xx, =X, =0 LPP involving two decision variables can easily be solved by graphical
to
Yes No No into the concepts of simplex method: a powerful technique
also provides an insight
decision variables
X X, =x, = 0 X solve the LPP involving three or more

follows:
major steps of graphical method
Yes No are as
No The
constraint as equation
Step 1: Consider each inequality
Problem 2: each one will geometrically
Step 2: Plot each equation on
the graph (xy plane), since
Find all basic solutions for the system of simultaneous equations
Tepresenta straight line.
2x, +3x,+ 4x, 5 =

The feasible region is the intersection


Step 3: Identify the feasible region (solution space). and is
the constraints of the problem
3x+ 4x,t 5x, =6 OT all the regions (common region) represented by
constraint is with 'S Sign then the
n = number of variables = 3 and restricted to the first quadrant only. If
the inequality
Solution: here of variables) is
the first quadrant (due to non negativity
m = number o f equations = 2 Tegion below the line lying in above the line
constraint is with ' 2 sign, then the region
Considered. If the inequality the equation ot
and then point on the line will satisfy
is considered. Every
t h e maximum possible number of basic solutions will be
ying in the first quadrant (intersection region) will satisfy all the constraints
iine. The points in c o m m o n region
tne called the feasible region.
nCmc3; n - m = 1, m = 2.
Stmultaneously. This common region is
= 0 X , 1 , X, =2
when Shade the feasible region
3/2 Step 4: The feasible region obtained in step 3 may be bounded or unbounded.
when X-2, x, = corner points of the feasible region.
Compute the coordinates of all the
when X, =0
x,-2,x, =3 function at each corner point determined in step
clearly, only three basic solutions exist, which are non-degenerate and infeastble also Step 5: Find the value of the objective
the value ot th

tep 6: Select the corner point that optimizes (maximize or minimizes)


00jective function. It gives the optimum feasible solution.
3.18 Linear Programming and Transportation Problem 3.19
Optimization Techniques

Remark 1: The above method is known as Search Approach method


Remark 2: Another method known as
Iso-profit or Iso-cost approach, involves the
following steps:
First
00M(O,1000) 1500R(0,1500)
a.
foursteps are same as in the search approach. In the fifth step we choose a
convenient profit (cost) and draw iso-profit (iso-cost) line so that it falls within the
1000

feasible region. 1000


500
b. Move this iso-profit line
parallel to itself farther (closer) from (to) the origin 500 A(1500,0)
L(2000.0)
C. Identify the optimum solution as the coordinates of that point
the feasible region on
0 500 1000 1500 500 1000 1500
touched by the highest possible iso-profit line (or lower- possible iso-cost line) x,+2x,=2000 x,+x=1500
d. Compute the optimum feasible solution. Figure 1
Figure 2

Solved Problems 4
Problem 1: (Maximization problem)
Solve the following LPP graphically
Maximize Z 3x, + 5x, 1000
subject to constraints: D(0,600) X600
600
x,+2x, s 2000
x,tx,S 1500 500

X 600
and x,20, x, 2 0 5001000 1500 2000
Figure 3
Solution:
2
combining the figurel, figure
Step 1: Considering each inequality-constraint as equation Step 3: Identify the feasible region (solution space) by below the
are with S sign then the region
x,+2x, =2000 (1) and figure 3 together. Since all the inequalities
feasible region
x , t x = 1500 (2) lines in Ist Quadrant should be considered, while shading
obtained 4) which is a set of point satistying
(Figure
X600 (3) Acommon shaded area OABCD is
the inequality constraints:
Cartesian rectangular
Step 2: Now we plot each equation on the graph by considering the
will certainly
plane. Clearly, any point (x,, x,) in the positive quadrant
1500R
axis OXX, in the
satisfy non-negativity constraints:

2000, put x,
x,20, x,
=
2 0

0, find x, equation, mark this


=
2000 from this
1000 x, =20003
plot the line x, +2x,
=
To
L similarly put x, to find x,
0 = 1000 and mark another point M(0,
=
D = 600
point (2000, 0) as

1000). Now join the points L and M. this line


will represent the equation x, + 2x, 2000
=
600E
500
P lying on or below the line x, +2x, 2000
=

as shown in the figure 1. Clearly, any point


to plot the
2000 similar procedure is now adopted
will satisfy the inequality x, t 2x,
600 shown in figures 2 and 3 respectively. Any egler
other two lines: x,tx, 1500 and x,
= as =

1500 and x, 600 will also satisfy other two


=

point on or below the lines x, +x,


=

inequalities: x, x, S
t 1500 and x, s 600 respectively. O 500 1000 500 2000
10.0) Figure 4
Problem
3.21
Linear Programming and
Transportation
3.20 Optimization Techniques

x,+2x, S 2000, x, +
x, S 1500. x, s 600 and non negativity restrictions as x, 2 0, x, 20.
Hence any point in the shaded area (including its boundary) is feasible solution to the given
LPP feasibleregion
Step 4: Find the coordinates of the corner points of feasible region 0,A,B,C and D.
Already, we got O = (0, 0), A = (1500, 0), D (0, 600) Now, we have to get B and C.
where B 1s the ntersection point of the lines x, + x, = 1500, && x, t 2x, = 2000

on solving these two cquations B= (1000, S00)


And Cis the intersecthon point ofthe lines x, = 600 and x, + 2x,- 2000

on solving these two equations C = (800. 600)

Step 5: Finding the value of objective function at each of these corner points (extreme
points)
Extreme point Z 3x+ 5x, 8
(0, 0) 0

(1500, 0) 4500 follows:


The value of objective function at each of
extreme points can be evaluated as
B (1000, 500) 5500 Z 201, + 40x,
(800, 600) 4500 Extreme point
D (0, 600) 3000 (0, 18) 720
Step 6: Clearly Z = 5500 is maximum for the point B(1000, 500) which gives the required B ,6) 280
optimal solution. C (4.2) L60
Problem 2: (Minimization problem) D (12,0) 240
9
Solve the following LPP graphically feasible
point C(4, 2)Hence it is the optimal
t r

Hence, minimum value of Z


=
160 for the
Z 20x, + 40%, B
Minimize lo,
solution.
subject to the constraints Problem 3: Solve the following LPP graphically
36x+ 6x,2 108 Minimize Z-2x,t
3x,+ 12x, 2 36 subject to the constrants
20x,+ 10x, 2 100 Sx+10x,sS0
and 20
Solution: Consider OXX, plane (rectangular). As each point has the coordinates of the
type (x, x,), any point satisfying the conditions x, 20, x, 2 0 lies in the first quadrant
, S4
only. The constraints of the given problem are plotted as described earlier by treating and
them equations. All the constraints are having '2' sign, therefore the area above these
Solution: Considering a set of rectangular Cartesian axis OX X, in the plane, and any
lines represents the feasible region in respect ofthese constraints; any pointon the stratght
lines or in the region above these lines would satisfy the constraints. The feasible region point ( x,) which satisties the condition x,20 and x,20 lies in the first quadrant only.
totting the given constrants by treating them as equations, we can have the tolloWing
of the problem is shown in Figure. The coordinates of the (corner points) extreme ponts
of the feasible region
8raph:
are
A-(0, 18), B =(2, 6), C=(4, 2) and D (12, 0)
Problem 3.23
Transportation
Linear Programming and

3.22 Optimization Techniques


Z 2x + 4x,
Extreme point (x,, %,)
(0,0)
X, 4 (4,0) 8

D B (3, 1) 1O maximum

Feasible
C (0, 2.5) 1O
the m a x X i m u m value
region the line segment BC gives
As Zmax 10 at B and C, .. any point on solutions or multiple
Hence there exists an alternative optimal
(10) of thee objective function
optimal solutions in this
case.
solutions- that is there are
infinite number of
function is parallel to a binding
5x,+ 10x, = 50
Remark on alternative optima: When the objective
the optimal solution) the
4
x, +X =

constraint (1.e., a constraint that is satisfied as an equation by


at more than one solution point.
function will assume the same optimum value
The coordinates of the extreme points are: objective this type can have
For this reason they are called alternative optima. The problem of
A = (1. 0), B (10, 0), C= (2, 4), D = (0, 4) and E = (0, 1)
infinite number of such solution.
Values of the objective function at these five vertices are
Problem 5 (Infeasible solution):
solution.
Z 2,Z =20, Z =8, Z, =4 and Z, =1 When the constraints are not satisfied simultaneously
the LPP has no feasible
The minimum of Z is 1 which occurs at E (0, 1). Hence the optimal solution to the constraints of the ' s type.
This situation can never occur if all the
are
problem is Min Z = l at (0, 1)
Solve the following LPP graphically
Froblem 4 (Alternative optimal solution):
Maximize Z =x, X2
Solve the following LPP graphically
subject to the constraints
Maximize Z 2x, + 4x,
X,tX,Sl
subjected to x,+2x, 5 -3x, +x,23
x,tx,S4
and
and x,0, x, 20
X X,20 follows:
Solution: The problem can be shown graphically as follows: Solution: The problem can be shown graphically as

Now, the value of objective function at each of the extreme points is evaluated as
follows:
- 3 x+
, X

C(0, 2.5)

WWTEWWWW B(3, 1)

A4, 0)
%,+2%, 5
3.25
Problem
and Transportation
Linear Programming
3.24 Optimization Techniques

The individual shaded regions o f -3x, + X, = 3 and x t X, = 1 are not intersecting


Problem 7: Solve graphically
4 and x,, X, 2 0.
that is there is no common region, and hence there is no point (x,, x,) which can lie in both + S 1, 2x, +
2x, 2
the regions (satisfy both the constraints), there exists no solution to the given problem,.
Maximize Z =
3x, 2x, subject to x,
-

x,
follows:
as
Solution: The problem is represented graphically
Hence there is infeasible solution
Problem 6 (Unbounded solution):
When the values of the decision variables may be increased indefinitely violating any of
the constraints, the solution space (feasible region) is unbounded. The value of objective
function, in such cases, may increase (for maximization) or decrease (for minimization)
nde finitely. Thus, both the solution space and the objective function value are unbounded.
Solve the following LPP graphically
Maximize Z 6x, t
X
subject to the constraints 4
2x, X, 23
X X , 20

and , X, 0

Solution: The given LPP is presented graphically as follows:


there is point (x,. x,) which satisties both the c o n s t r a i n t s
Ciearly from the figure, no
in
has solution because the constraints are
Simultaneously. Hence the problem
no

consistent.

Problem 8: Solve the following LPP graphically


Maximize Z = x , + x, subject tox, - X, 2 0

-3x. X, 2 3 and x,, x, 20


+

of feasible solutions in this


Solution: The following graph shows that there is no region
case

The two extreme points ofthe feasible region are A(1, 1) and B(0, 3), and we observe
that the feasible region (solution space) is unbounded. The value of the objective function
at the extreme points A(1, 1) and B(0, 3) are 7 and 3 respectively,.
But there exist number of points in the feasible region for which the value of the
objective function is more than 7. For example, the point (3, 6) lies in the feasible region
and the objective function value at this point is 24, more than 7. Thus both the variables x,
and x, can be made arbitrarily large and the value of Z also increased. Hence, the problem
has an unbounded solution.

does not arise


ec tnere is no teasible solution. So the question of having optimal solulion
Problem 3.27
3.26 Linear Programming and Tran sporta tion
Optimization Techniques
to infinity on the
thisedge (-x, + 2x, 4)
= which is going
Problem 9: Solve the following LPP graphically Every p0int (x,, X,) lying on

Z 4 and is therefore an optimal solution.


right gives
=

Maximizze Z= Sx, +
3x,
Subject to 15 Problem 11: Solve the following LPP graphically
3x, +5x, =

0 and x,, x, 20
2x, subject to x, S 3, x, x,
+
-

S
Sx,+2x,= 10 Maximize Z =

-3x,
follows:
and XX,0 Solution: The given problem is represented as

Solution: The following figure shows the graphical representation of the given LPP:

(feasible region consists of


only one point A)

,+5x,
all the variables
does not necessarily imply that
In this problem, an unbounded solution constant
Z approaches infinity. Here the variable x, r e m a i n s
(20 can be made arbitrarily large as
Since there is only one solution point A 19' 19 there is nothing to be maximized. (shown in figure)
Hence, a problem of this kind is of no importance. Such problems can arise only when the LPP : Important geometric properties
known as the convexity
number of equations in the constraints is at least equal to the number of variables. If the has an important property,
1. The region of feasible solutions convexity
solutions is feasible, it is optimal. If it is not feasible, the problem has no solution. the feasible solution of the problem exists
property in geometry provided them that is they aresolids.
feasible solutions has no holes in
-

Problem 10: Solve graphically means that region of

Maximize Z=-x, + 2x, subject to -x, + x, S 1 and they have no cuts on the boundary.

-x, + 2x, 4 and x,, x, 20 11. The boundaries of the regions are lines orplanes
various corners.
There are c o r n e r s or extreme points on the boundary, and edges joining
Solution: The problem is represented graphically as follows 11.
value of Z
function can be represented by a line or a plane for any fixed
IV The objective solutions
feasible solutions will be an optimal
V. At least one corner of the region of finite
the m a x i m u m or minimum value
of objective function (Z) 1s
when ever
other than corners that are
Vi. If theoptimal solution is not unique, there are points
least one corner is optimal.
optimal but in any case at
function can be made arbitrarily
VIl. The different situation is found when the objective
that case.
large of course, no corner is optimal in
2 x ,
=4

Solved problems 5
Problem 1:
Find all the basic solutions to the following problem
Maximize Z x t 3x, + 3x,
Problem
3.29
and Transportation
3.28 Linear Programming
Optimization Techniques

Subject to X,t2x, + 3x, = 4 R, -2R, 0 5 617 7


2x,+ 3x, + 5x, = 7
01-4 -13-11
Also find which of the basiC solutions are RR, -5R, LO 0 -8 -24-24
1. basic feasible
11. non-degenerate basic feasible R,R,+ R,
canonical formas:
i. optimal basic feasible on a,, to obtain the required
Finally pivot
Solution: n = 3 and m = 2, n - m =1
R3
We can have 3c,= 3 different Basic solutions R 0 0 1 2-
8
S No Basic Non Values of Z feasible?Non degenerate? Optimal R R, -SR, 0 0 1 3
Basic Basic variables|
feasible R, R, + 4R,
have the canonical form
X,, X X0x, 2, x, =1 5 Yes Yes Yes From this matrix, we can

X X, =0x, 1, X, =1 Yes Yes No X,tX 2


(2)
X X,X,0X,-1,x, =23 No Yes No
From the above table, the first two solutions are x+3x3
non-degenerate feasible solutions and
the third is non-degenerate and infeasible, the first solution is the optimal one. and x, = 2 - x

The optimal solution is


x,=1tX,
ms5 at (2, 1, 0).
Problem 2: Find all the basic solutions x,3 3x,
corresponding to the system of Equations the independent
2x, + 3x, -2x, - 7x,=1 If Eqs. (1) are the constraints of a LPP, the solution obtained by setting
variable equal to zero is called a basic solution.
x,tXtx, +3x, =6 (1)
n the present case, the basic solution is given by
X X , t x , + 5x, = 4
=3 (basic variables) andx, =0 (non basic variable).
Solution: First we have to reduce the system of equations into a canonical form with x, X,2, X,= 1, x,
x, and x as basic variables. it is BFS.
Since this basic solution has all x, 2 0 G =1, 2, 3, 4). a

For easy calculations, consider all the coefficients of variables, and constants as a
1I we want to move to a solution we can proceed from the canonical
neighboring Basic
matrix: Irom given by equation (2). Thus if a canonical form interms of the variables x,, X, and x

place original
in of the basic variable X
required, we have to bring x, into the basis
1SHence we pivot on a, in equation (2). This gives the desired canonical form as
4J
R3
For this, first we pivot on the element a, 2 to obtain:
X, X,1 by R,>

R -1-1 , 2 R R, - R
R, R, -R, 02 X, X=1 R, R,+R,
R,R, R, Lo-2 Ihis canonical system gives the solution of x,, X, and X, in terms of X, as

then to obtain
pivot on
a,
3.30 Transportation Problem
3.31
Optimization Techniques Linear Programming and

3.1.5 The need of simplex method


of LPPs in two
discussed the geometrical interpretation
In the earlier part of LP, we described mathematically
feasible solutions to a LPP
dimensions we observed that a set of
2-, (3)
formed a convex region or convex set (if
there was a feasible solution).
Further more
the convex
point of
solution existed, at least one corner or extreme
when an optimal
region was optimal.
and the corresponding basic solution is given by basic
solution we might attempt to find all
To determine an optimal basic feasible
x, =1, x, =2, x, = 1 (basic variables)
the optimal value of the objective function.
solutions and then select the one which yields done.
x, =0 (non basic variable) and hence in theory this could be
There is only finite number of basic solutions,
because the number of basic solutions
This basic solution is a BFS. However, this procedure is extremely ineffective, 10 and
variables are increased For example if n
=

increase very rapidly as the number of


Ifwe want to into the
to the next basic solution with
move
x, X, and x, basic variables, as we
have 252 basic solutions. Usually, we do not have to inspect
all these basic
have to bring x, current basis in place of
x, Thus have to pivot a, " we in m =6, we
this
solutions since many of them will be infeasible,
however, for large values of n and m,
Equations (3) scheme
by one. Hence we need a computational
IS still a very large number to inspect one
R3R X,tX3 which corresponds to a optimal value of
f
that examines a sequence of BFSs, each of
R2
R,R,3 X,+3x,=6 (4) until an optimum is reached.

LPP. it will be useful to review


Before studying the most general method of solving
a
R2
RR,+ XX, - 1 the methods of solving a system of linear equations.
and x, 3-x, in unknowns
Consider the following system of n equations n

X 6-3x,
- 1 + X, a X,a2 X, t . a X , D,
.. the basic solution is:x =
3, x, =
6, x =
-1 (basic variable) a X, tan X,t a, X, =b,
b, (1)
ta, X,
=
t
x, =0 (non basic variable) a X,t a2 X,
Since all the x, are not nonnegative, this basic solution is not feasible.
Now, Finally, to obtain the canonical form interms of the basic variables
x,, X, and x
we pivot on
a,," to bring x, into the current basis in place of x,
This gives
solving method of
R R Assuming that this equations possesses a unique solution, a
set of
X,tx,3 the equations to a form known as canonical form by
the system consists of
reducing
R,R,-3R, x,-3x,-3 using elementary row operations, the system of above equations
can be reduced t0 a

R,R, + R, xtx, = 2
convenient equivalent form, such that, a particular variable will be
eliminated from all
all
This canonical form gives the solution for This procedure of eliminating a particular variable from
x,, x, and x, interms of x, as cquations but, one equation.
the
one cquation is called a pivot operation. The system of equations produced by
%3-, Dut, That is, the
%,=-3 + 3x, Pivot operation have exactly the same solution as the original set of Equations
vector X that satisfies (1) satisfies equation (2) and vice versa.
x2-, outn
and the corresponding basic solution is Now, considering the reduced form (canonical form), obtained after carry1ng
x, 3 , X,-3, x, = 2 (basie variables) pivot operations:

and x 0 (non basic variable)


This basic solution can also be seen to be infeasible due to the negative value for x,
Problem
3.41
Transportation
Linear Programming
and
3.40 Optimization Techniques
can be m o r e
the previous problem.
the simplex procedure for
By using the short cuts,
represented by the following single table
comveniently Min.ratio
Basic
variables
C
Z-11 0

Now, the improved simplex table: 5-Min.ratio


0
BV X,
31 04
0 Z-C,X, -0
, min ratio

3-1 -
Z-CX,1 o ,--0 Z-C,X,
The improved solution from this table is: -
x,3,x,1, s,0, s,0 and 7,11 3
Also, using the formula ACX-C, clearly all
that is A 20vj
A, s ane non-negative
-All A, z0
Hence the optimum solution is 8,0 Z-CX,11
,3X,1.2,11 Thas,the optimal solution is obtained as: x,-3. x, 1,2 11
Shortcuts for quick solution by simplex method
i In the first iteration only, since
need
As, are same as-Cs (since C,0). so there mo
of caleulating them separately by using the formual A-C,X-C
is
Solved problems 6
Problem 1: Solve the following LPP by simplex method 2 7
Mark min(A) by T which at once indicates the column X, needed for computing the Maximize Z-15x,+ 6x,+9x, + 2x,
minimum ratio subject to 2x, X, 5x,+ 6x, 20 2-+2
3x,x,+3x, +25%,s 24 1 - 2 2
Key element
and the left directed
is found

arrow
at the

(-)
place where

of minimum
the

ratio
upward directed arrow

intersect each
of

other
min

in
3
the
and
7x,

g
X 70

gX, 20
6
Simplex table
Solution:
-O-2
Key element indicates that the current table must be trans formed in such a way that
the key element becomes 1 and all other clements in that column become zero
Writing the given LPP in the standard form by introducing non negative slack vartables

Since As corresponding to unit column vectors are always zero. thère is no need of S,5, and s
calculating them.
While transforming the table by row operations the value ofZand corresponding Aj's
ane also computed at the same tine.
Maximize
subject to
Z 15x, + 6x, + 9x, t 2x, + 0s, + 0s, + 0s,

2x+x,+ 5x, +6x, + s, =


20
-1-2
3.42 Problem 3.43
Optimization Techniques Linear Programming and Transporta tion

3x,+x,+ 3x, + 25x, + s,


=
24 The optimal solution is given by
Tx, +Ox, +Ox, t x, +s, = 70

and 2 0 X,12
X X Xg S,, S, S,
Now the IBFS is (x,=X, =

X,
= =
0 non basic)
x, =

x,
s , 2 0 , s = 24, s, = 70 basic x 0 and Max Z 132

The Problem 2: Use simplex method to solve


simplex procedure is given in the single table by using shortcuts.
C 15 6 2 0 0
Minimize Z=x, -

3x, +2x,
Basic Min.ratio
subject to 3x,-X,+ 2x, 7
variables
, S, -2x, + 4x,S 12
10
0 20 2 20 -4x, +3x, +8x,
0
2-10 and

0 24 24 Solution:
1 3 25 0 1 0 8
3 Since the given objective function is of minimization type, we can convert it into
min ratio0
maximization type
0 70 70
S 0 0 0 10 Min Z Max (-Z)
7
Z0 -151 6 -9 2 0
Maximize (-Z) = Maximize Z*=-x, t 3x,-23,
0

subject to 3x,-x,+ 2x, $7


S 3 0 12*| -2x, + 4x, S 12
m i n ratio -4x, + 3x, +8x, S 10
8
15 1 5 0
and XXX,0
-24 slack variables
and writing the given LPP in standard form by introducing non negative
172
14 -7 0 SS, and s
3
Maximize Z-x+ 3x, -2x,+ s,= 7
Z=120 6 123 04 0
subject to -2x, + 4x,ts, = 12
6 12 9 32 3
57
4x, + 3x, t+8x,+ S, =10
15 2 and
X Xy Xg S, Sg, S, 10
0 42 0 14 132 -7 .TheIBFS S7, s, = 12, s, = 10
Z= 132 0 15 91 (xX,X, = 0, non basic)
3 0 All A, 2

Since all A, 20, the current BFS is


optimal
3.44 Optimiza tion Techniques Problem 3.45
Linear Programming and Transportation

C 3 0 0 0 Solution:
BV
C X X, X S, S, Min.ratio Writing the given LPP in standard form
Max Z 4000x, + 2000x, + 5000x, + 0s, +0s, + 0s,
by introducing slack variables s, S, and s,

0 3 0
12x,+ 7x,+9x, + s, 1260
=
subject to
12 -2 0 3 min ratio 22x, + 18x, + 16x, + s, 19008
=

2x, + 4x, + 3x, t s, 396


=

0 10 4 3 8 0 0 1
103.33 and XXgXyS,, S,, S, 20
Z* = 0
-31 2 0 0 0
The starting solution and subsequent simplex tables are:
0 10 20
4 4000 2000 5000 0 0 0
5 C,
Basic Min.ratio
0 0
variables C, X, S

0 1260
0 0 1260 12 7 9 0 0
9
19008
Z* = 9
0 0 0 0 19008 22 18 16 0
16
396
4 1 0
10
0 0 396 4 0

3 0
4000 2000 50001 0 0 0 A
10 0 12
0 72 0
0 11 0 0 10 10 6
S 0 16896 0 149i
Z-110 all A20
5000 132 198

since all A 20, the current BFS is optimal 2600 14000 5000
Z 660000 0 0
3 3 3
Theoptimalsolution:Z11,X,-4, x, =5, X, - 0
and Min Z = Max (-Z) = -11, x, =4, x, = 5, X, =0
4000 12

Problem 3: Maximize Z 4000 x, +2000 x, + 5000 x, 55 7


16760 0
9
subject to 12x, + 7x, +9x, s 1260
S000 124 0 0
22x+18x,+16x, < 19008 9 9

2x+ 4x, + 3x, 396 Z= 668000 0


37000
0
1000
0
4000
all A20
9 9
and X,XX,0
.46 Optimization Techniques Linear Programming and Transportation 3.47
Problem

sinceall A 20, the optimal solution is given by 20 6 8 0 0 0

X 12 Basic Min ratio

X,0 variables C X, X, X, SS, S,


x,= 124 and 250
Zmax668000 0 250 8 1 0 0
8
Problem 4: Solve the following LPP by simplex method 150
0 150 4 0 0
4
Maximize Z =
20x, +
6x, +
8x,
subject to 8x, +2x, + 3x, S 250 S 0 50 0 0 0 1 25
min ratio
4x, +3x, s 150 Z 0 -201 6 8 0 0 0

2x, tx, S 50 50
0 50 0 2 0 4
and

Solution: Writing the given LPP in the standard form by introducing non negative slack S2 0 50 0 -2 1 Min ratio
variables S,, S, and S,:
20 25 0 0 0
Maximize Z* 20x, +
6x, + 8x, + Os, +
0s, +
0s,
subject to
Z 500 0 -61 0 0 0
8x,+2x, + 3x, t s, =
250
50
50
4x, + 3x, + s, 150 = 0

2x, t x,t s, 50 =

0 1 0
and Xs S,, S,, S, 0
X,,X
The IBFS is given by s, = 250, s, = 150, s, = 50 (x, = x, = x, = 0, non basic)
20 25 0 0 50*
Considering the simplex procedure in a single table:
Z 600 0T 0 -2 6

0 0 0 0

4
6 50 0 0 0
3
8 50 0 1 0 0 1
Z 700 0 0 0 2 8 all A 20

Sinceall A 20, the current BFS is optimal

The optimal solution is x,=0


X,50

x, 50 and Zm 700
Problem 5: Solve the following LPP using simplex method
Maximize Z 10x, + 15x, + 20x,
subiect to 2x + 4x. +6x, 2 4
3.48 3.49
Optimization Techniques Transportation Problem
Linear Programming and

3x, +9x, + 6x, 30


3x, +Os, 8
=

and
subject to 2x +

2x,+Sx,+ Os, 10
=

Solution: The standard form of the given LPP is and X X X, S,, S,, S,20.
3x, + 2x, t 4x, +0s, 15
=

Maximize Z 10x, +15x, + 20x, +Os, + Os, =


0, non basic)
10, s, 15 and (x, =x, x,
=

the IBFS is s,
=
8. s, = =

subject to 2x, + 4x, + 6x, + s, = 24 tables are:


the initial simplex table and subsequent
3x+9x, + 6x, + s, = 30
C 3 5
and 2 0
X, Xg X S, S,
the initial simplex table and subsequent simplex tables are: Basic Min.ratio
S
C, 10 15 20 0 0 variables C X,X X, X
= min ratio
Basic Min 2 0
Variables C, X, X, x, ratio 10
10 2 0
2
24
24 2
4 min ratio 15
15 4 0 0 1
0 30 3 9 6 0
Z 0 -3 -st-4 0 0 0 A
Z 0 -10 -15 -201 0 0

20 1 0
6 G-12 -/5 min ratio
0
5 0

0 6 0 -1 6 min ratio
-10
0 0 4 0 14
Z = 80
0
0 0 40
Z- 0 -41 0 0

20 2 0 -1 1 5
0
10 6 (G)
Z 100 15 0 0 all A 20 4
5
Since allA 20, the current BFS is optimal 89 41
The optimal solution is given by S 0 0 0
x 6 , X=0, x, = 2 and Z m 1 0 0
256 7 4
Problem 6: Solve the following LPP by simplex method 1 15 15
0
15 0
Maximize Z 3x, t Sx, + 4x,=

50 5 10
subject to 2x, +3x, S8 5
41
0 1 0
41
2x, + 5x, s 10
2
3x, + 2x, + 4x, S 15 0
and X,X X 20. 41 41
Solution: Writing the given LPP in the standard form by introducing non negative slack 89
0 0
variables S,, S, and 41
S,
Z 765
Maximize -3x, 5x, 4x, Os, 0s, 0s,
+ + + + +
Z 0 0 0
24
all A20
41 41 4
Problem 3.79
Linear Programming and Tran sportation
3-78 Optimizeticn Techmiçues
MAXIMIZATION
SIMPLEX ALGORITHM FOR
min {A, 3}- 1hen
select inceming vecter, find such value of Kfor which A
=
e
he vector coming into the basis matrix wnll be X
matrix will be X, it
l o s e e t entgoing vecter". The vector going out of the basis
e deter e the s t i t rby the minim a t o rule Express the LPP in
Formulate the given
the standard for
problem as LPP

or predetermined value of k

transformatioa Obtain an lBFS to


ep Now onstrectthe nettproremeat table by using the simple matrix sLPP
us

Step 9: Now retm to step 6, then go to the steps S and 9 if necessary. This procedure is
Construct the startng
rpetedl e pt aBFS is obained.
sinplex teble wih the
We bare sem that h e r programing is a pomerful techmique for solving the problem above IBS

ofallocat i e s o u r c e s r g competing activities as well as other problems havin


a s i i =htical e We have seen how a real life problem can be formulated Update the simplex Compute he net
s a LFP. Also we bve seen some limitations in the construction of LPP which leads to table by appropnate
matrr ensformatbion
evauatons
ehecbounded soluion or o solutio
pivobn
F e t e he set of feasble solutions is a comver set. In fact we use the result that the
caerer set of feasble solutioms of a LPP a s at least one extreme point and amost a
Remove the e v Souson
e be o f e t e p o s Tbe objective fection is optimized at one of the extreme vereble rom the besis der test s
p u s I S g a cucvet e t wiich is defied by a set of linear constraints in an n- and troduce the
ls any o
entern9 venebe
asto
s sgece e e reqred to detemne subset of points iron this convex set Find key element
for wiich he robjective fecin is optimized
We ere given he working rules to solve a LPP using the simplex algorithm. The Choese e n Get the vector x
algar e i s tzt the give roblea should benhe standard formn. The initialization raion he Basi
the
eeds he existence ad iientifcatio of an 1BFS. This necessitated the introduction of verable corespoi to Entening
hat
basis so
min
ves the bess
E b l e s called sack variables, supls vziables and if these do not serve the purpose
then the i f a l rables. An iteratire process involving the construction of a series of
tbles naerhnncal way moald ultiately lead to the solution oftheproblem. To deal Soltion can be
proved. calculate theNo Coresponding to any
w2nd=comstraiets, the Big-M=ebod is explained. Exceptional cases like degeneracy Ac0, are al he Solons
enbounded
Altermete optime etc_ zre discssed in detzil with worked out examples. ratio o) elements of enterins
ector xs0 step

Bits
Linear programming
variables are introduced make < type in
to
equahtes
Ans: slack
A system with m equations and n vanables has at most basic solunons.
Ans: n

You might also like