Documente Academic
Documente Profesional
Documente Cultură
1. Introduction
1.1. Definition of Operations Research
There have been various definitions for Operations Research (O.R.) like applied
decision making, quantitative common sense, making of economic decision, etc.
Here we pick one of the most appropriate;
Definition: Operations research
Operations research is the application of up to date scientific methods,
techniques by inter-disciplinary teams to problems involving control of organized systems so as to provide solutions which best serve the purposes of
SMA 2343
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
Operations Research 1
Reduces complexity:- Simplifies the work of executives especially due to prior
experimentations.
Operations Research is therefore both a science and an art.
1.3. Methodology of Operations Research
(a) Problem identification: Recognition that a problem exists is very
important in any management decision-making process, but in
practice its timing may be critical (resolve an existing problem
or to forestall a predicted problem).
(b) Formulating the problem: Once it becomes apparent that a probformulated in terms of;
SMA 2343
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
Operations Research 1
(e) Testing the model validity: The solution values of the model, obtained at solutions stage are then tested against actual observations. In other words, effort is made to test the validity of the
model used. A model is supposed to be valid if it can give reliable
prediction of the performance of the system represented through
the model. In effect, performance of the model must be compared
with the policy or procedure that it is meant to replace.
(f) Controlling the solution: This step of an O.R establishes control
over the solution by proper feed-back of information on variables
which might have deviated significantly.
SMA 2343
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
Operations Research 1
2. Classification of Problems in O.R.
Broadly speaking, problems in O.R. can be categorised; Some of the categories
are;
1. Allocation: Allocation problems involve the allocation of resources to
activities in such a manner that some measure of effectiveness is optimized. E.g. Jobs to applicants, Money to investment projects
2. Replacement: Replacement problems are concerned with situations
that arise when some items (such as machines, electric light bulbs, etc.)
need replacement because the same may deteriorate with time or may
break down completely or may become out-of-date due to new develAll rights reserved
SMA 2343
opments. Replacement problems thus occur when one must decide the
optimal time to replace equipment for one reasons or the other.
3. Sequencing: Sequencing problems are the problems concerned with
placing items in a certain sequence or order for service. For instance,
N-jobs requiring different amounts of time on different machines must
each be processed on M-machines in the same order with no passing
between machines, then the question: How should the jobs be ordered
for processing to minimize the total time to process all of the jobs on all
of the machines constitutes an example of a sequencing problem.
4. Routing: Routing problems are problems related to finding the optimal
route from an origin to a destination when a number of alternative routes
are available. For example, a salesman may wish to visit each of Ncities once and only once before returning to his headquarter, then his
problem is: In what order should he visit the cities so that the overall
distance travelled is minimized? Such a problem is referred to as a
routing problem.
5. Inventory: Inventory problems are problems with regard to holding
or storing resources. The decisions required generally entail the determination of how much of a resource to acquire or when to acquire it.
The problem of deciding how much of a certain commodity to hold in
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
Operations Research 1
inventory is one of real concern to business and industrial houses. Inventory problem is the problem to determine the level of inventory that
will optimize the measure of effectiveness.
6. Queuing: Queuing problems or what are known as waiting-line problems are problems that involve waiting for service, Queuing problems
encircle us from the time we rise in the morning until we retire at night
In business world several types of interruptions occur: facilities break
down and require repair, power failures occur, workers or the needed
material do not show up where and when expected. Allocation of facilities Considering such interruptions be done and to do so means solving
SMA 2343
a queuing problem.
7.
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
Operations Research 1
(e) Simulation
(f) Game theory
(g) Decision theory
SMA 2343
inquiry
4. (2mks) Under which category of OR problems is Decision theory?
End Quiz
Marks:
Correct
Percent:
Comments:
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
Operations Research 1
3. Linear Programming
Linear Programming deals with problems in which linear functions are to be
optimized (maximized or minimized) subject to constraints which are usually
specified by linear inequalities, linear equations and to the condition that all
the variables must assume negative values.
The general formulation of linear programming problems is as follows.
Optimize the objective function
Z = c1 x1 + c2 x2 + ... + cn xn
Subject to the linear constraints
SMA 2343
x2 0,
x3 0, ...xn 0
There are many problems in engineering management and social sciences that
can be formulated this way.
In matrix form we can write as
~
Optimize Z = ~c> X
~
~
subject to AX{,
=, }B,
~ 0.
X
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
Operations Research 1
Definition: Optimal feasible solution
Optimal feasible solution is the set of decision variables which is feasible
and optimizes the objective function.
In order to obtain the Optimal feasible solution, we need to have some concepts
from linear algebra.
3.1. Convex analysis
Definition: A line in Rn or n-dimensional space
The set
SMA 2343
(0, 1)
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
Operations Research 1
If H is included in the spaces s1 and s2 , then we have the closed half spaces.
sc1 = {X|a> X } and sc2 = {X|a> X }
Example. Show that a closed half space sc1 = {X|a> X } is a convex set.
Let X 1 and X 2 be two points in sc1 then a> X 1 and a> X 2
then any point X on the segment joining x1 and x2 is given by
X = X 1 + (1 )X 2
(0, 1)
SMA 2343
and i 0 i
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
i=1
Operations Research 1
Theorem 3.2 A set of convex combinations of the points x1 , x2 , ..., xn C
where C in convex is also convex. Let C be the set of all convex combinations
n
P
of x1 , x2 , ..., xn , let Y1 and Y2 be two points in C, then Y1 =
i xi and
Y2 =
n
P
i=1
i xi
i =
i =1
i=1
(0, 1)
x0 = Y1 + (1 )Y2
n
n
P
P
= i xi + (1 ) i xi
i=1
n
P
i=1
(i + (1 )i )xi =
i=1
i xi
i=1
i = i + (1 )i
SMA 2343
n
P
and (0, 1)
n
P
i + (1 )i
P
P
= i +(1 ) i
i =
i=1
=+1=1
Thus x0 =
n
P
i xi
wherei 0and
i=1
n
P
i=1
Example. Prove that any point inside the triangle is a convex combination
of x1 , x2 andx3 Let x0 be any point inside the triangle. By definition x0 =
x1 + (1 )Y
(0, 1)
But
Y = x2 + (1 )x3 (on line x2 , x3 )
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
(0, 1)
10
Operations Research 1
Therefore
x0 = x1 + (1 )[x1 x2 + (1 1 )x3 ]
= x1 + 1 x2 + (1 1 )x3 1 x2 (1 1 )x3
= x1 + ( 1 )x2 + (1 1 + 1 )x3 .
3
P
x0 =
k xk
k=1
where
1 =
2 = 1
3 = 1 + 1
x0 =
3
P
SMA 2343
k=1
3
P
k = 1 x0
k=1
(0, 1).
We note that the extreme points are actually the vertices or corner points.
Example
Consider the figure
figure
It is not possible to locate two distinct points in or on the above figures(Convex
sets) with the property that the line joining these points will include the extreme points (E) and be included in the convex set.
Theorem 3.3 The convex set of feasible solution of linear programming problem has at least one extreme point and at most a finite number of extreme points.
The objective function is optimized at one of the extreme points.
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
11
Operations Research 1
Note that the convex set in a linear programming problem is the constraint
set.
3.4. Optimal solutions to linear programming (LP) problems
In Linear Programming problem we are required to optimize the objective
function
p = C0 X
under a constants set. Suppose that R is the constraint set of the given
problem, then the optional solution is an extreme point of R
Theorem 3.1 A basic feasible solution of a linear programming problems is
All rights reserved
SMA 2343
Graphical method
This involves representing all the constraints on a graph. The region of intersection represents the feasible region and by theorem (3.1), one of the extreme
points gives the optimal solution.
Example. The manager of a theatre which has a capacity of 300 seats sells
tickets to children and adults at Shs.10 and Shs.50 respectively. To cover his
rental expenses, he has to take at least Shs. 2500 for each show.
It is the companys policy to have at least 100 seats spared for children.
Formulate this problem for profit maximization and solve it graphically.
Solution: Let x and y be the number of 10/= and 50/= seats occupied respectively. Then we need to;
Maximize 10x + 50y subject to;
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
12
Operations Research 1
x + y 6 300 (restriction on the hall capacity)
10x + 50y > 2500 (restriction on total collection)
x 100 (restriction on seats for children)
y0
SMA 2343
Graph
Figure
1:
One
point
at
the
vertex
profit.
At point;, Testing for optimal
A(100, 30), z = 10(100) + 50(30) = 2500
(corner)
solution.
maximizes
At
the
point
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
13
Operations Research 1
Let x and y be the number of ha on maize an beans respectively then, we
have
Maximize p = 30x + 40y subject to
x + y 50 (Constraint on land)
4x + 2y 150 (Constraint laborers)
50x + 100y 4500 (Constraint on capital)
x, y 0 (Non-negativity condition on decision variables)
SMA 2343
Figure
2:
Testing
for
A(0, 0), p = 30(0) + 40(0) = 0,
optimal
solution.
At
point
B(0, 45), p = 30(0) + 40(45) = 1800
C(10, 40), p = 30(10) + 40(40) = 1900 B(25, 25), p = 30(25) + 40(25) = 1750
(37.5, 0), p = 30(37.5) + 40(0) = 1125
The maximum occurs at x = 10, y = 40, The farmer should therefore
cultivate 10 ha of maize and 40 ha of beans order to make a maximum profit
of $1900.
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
14
Operations Research 1
should give a better Solution as you increase or decrease the decision variables.
Definition: Slack variable
Slack variable a non-negative variable which when added to a less that or
equal to inequality makes it an equation.
Example.
x+y 7
but
x + y + s1 = 7
SMA 2343
Let x and
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
15
Operations Research 1
SMA 2343
s1
s2
s3
Solution
s1
50
s2
75
s3
90
-30
-40
Tableau #2
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
16
Operations Research 1
Basis x
s1
s2
s3
Solution
s1
1/2
-1/2
s2
3/2
-1/2
30
1/2
1/2
45
-10
20
1800
SMA 2343
s1
s2
s3
Solution
-1
10
s2
-3
15
-1
40
20
10
1900
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
17
Operations Research 1
A
M1
600
M2
500
M3
800
SMA 2343
<= 600
3x +4y
<= 500
4x +6y
<= 800
x, y >=0
The final table is #3 given by
Basis x
s1
s2
s3
Solution
2/5
-1/5
140
-3/10
2/5
20
s3
1/5
-8/5
120
14
8200
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
18
Operations Research 1
2. A company produces two types of leather belts; say A and B. Belt A is
of superior quality and belt B is inferior. Profits on the two are sh40
and shs 30 per belt respectively. Each belt of type A requires twice as
much time as required by a belt of type B. If all belts were of type B, the
company could produce 1,000 belts per day. But the supply of leather
is sufficient only for 800 belts per day. Belt A requires a fancy buckle
and only 400 of them are available per day. For belt B only 700 buckles
are available per day. How should the company manufacture two types
of belts in order to have a maximum overall profit? (solve it graphically
and using simplex method).
SMA 2343
3. The manager of a hotel has sufficient money to buy a total of 100 crates
of soft drink of types A and B. He wants to buy at least twice as many
crates of type A as type B. He wants to buy maximum 80 crates of type
A and at least 10 crates of type B. Taking X to be the number of type
A crates and Y that of type B, write down all the inequalities based on
these facts. Show these inequalities on a graph and outline the region
in which (X, Y) must lie. The profit on a crate of type A is Shs. 60
and that of a crate of type B is Shs. 40. Find the number of crates of
each type that he should buy to make maximum profit and calculate this
maximum profit. (solve it graphically and using simplex method).
4. A chemical firm has 160 litres of solution A, 110 litres of solution B and
150 litres of solution C. To prepare a bottle of syrup X, 200ml of solution
A, 100ml of solution B and 100ml of solution C are needed. For a bottle
of syrup Y, 100ml of A, 200ml of B and 300ml of C are needed. Syrup
X sells at Shs. 60 per bottle and Syrup Y sells at Shs. 100 per bottle.
How many bottles of each type of Syrup should the firm make in order
to obtain the maximum amount of money? (solve it graphically).
5. A farmer requires 10, 12 and 12 units of chemicals A, B and C respectively for his garden. A liquid product contains 5, 2 and 1 units of A, B
and C respectively per bottle. A dry product contains 1, 2 and 4 units
of A, B and C per carton while a paste product contains 1, 3 and 1 units
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
19
Operations Research 1
of A, B and C per jar. If the liquid product sells at shs 30 per bottle,
the dry product sells at shs 20 per carton and the paste product at Ksh
25 per jar, How many of each should he purchase in order to minimize
the cost and meet the requirements? (Formulate and dont solve).
6. A farmer has 70 hectares of land available for growing maize and beans.
The cost of growing 1 ha of maize is $30 and the cost for growing 1 ha
of beans is $20 and the farmer has only $1800 available. The labor per
ha is 2 man days for maize and 4 man days for beans and a total of
240 man days of labour are available. If he makes a profit of $800 for 1
ha of beans and $700 for 1 ha of maize, formulate the underlying linear
SMA 2343
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
20
Operations Research 1
the number of newspaper adverts must be no more than 2.5 times the
number of combined radio and TV adverts.Formulate the problem if the
company wishes to maximize the number of customers exposed to the
advertising campaign. (solve it graphically).
A company publishing textbooks is planning its production of the next book
scheduled to be printed. The book will be published in both paperback and
hardback format. The paperback sells for 10 per copy and costs 5 to produce and market. The hardback sells for 20 and costs 17. Market research
has indicated that total sales of the book are unlikely to exceed 10000 copies.
Of these at least 4000 are expected to be in paperback format with at least
SMA 2343
2000 hardback. On the other hand, the company does not expect to sell more
than 4000 hardback copies. In addition, there are potential problems involved
in producing the paperback edition. The printing equipment is needed for
other paperback books and is available for printing this book for a period of
only 5000 hours. Each paperback takes 40 minutes to produce.
1. Formulate this problem and solve it assuming the company wishes to
maximize profit.(solve it graphically).
2. Assuming the company wishes to maximize revenue from sales reformulate this problem and solve it and computes the profit the company
would earn. (solve it graphically)
3. The production manager is strongly arguing that production should be
determined by costs. Reformulate the problem in terms of cost minimization solve it . (solve it graphically).
4. Which of the three alternative objective functions do you think is most
appropriate?
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
21
Operations Research 1
Quiz Solve the LP below using dual simplex method
M aximize z = 150x + 250y + 450z
subject to
2x + y + 3z 80
3x + y + 4z 60
x + 2y + 5z 100
x 0, y 0 , z 0
Optimal Solution: p =
; x=
, y=
, z=
NOTE
SMA 2343
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
22
Operations Research 1
1. If no slack variable do not appear in solution set, then all the resources
have been exhausted.
2. The solution quantity corresponding to si is the amount of resources
which were not utilized/abundant resource.
3. Values at profit row corresponding to Slack variables gives the marginal
value products of the corresponding resources (per unit contribution to
profit). They are also referred to as dual prices and in economics
shadow prices or imputed costs. They represent the worth per unit of a
resource bi i = 1, 2...m.
In the farmers problem. s2 = 15 indicates that (15 2) = 30 laborers
All rights reserved
SMA 2343
will be un utilized.
4. The p row coefficients of s1 , s2 and s3 are 20, 0 and 10 respectively
implying that;
Increasing land by 1 ha increases profit by $20
Increasing the number of employees has no effect on profit
Increasing the number of capital by $1 increases the profit by
$10/50
The farmer should not pay more than $20 for every additional 1 ha hired.
It would not make sense to put more money in form of capital.
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
23
Operations Research 1
Exercise 3. A carpenter makes boxes, tables and chairs. The profit contributions of the three products are $20, $30 and $10 respectively. The carpenter
can afford to spend up to 40 hours per week working and takes two hours to
make a box, six hours to make a table and two hours to make a chair. Customer demand requires that he makes at most a third as many boxes as the
total number of chairs and tables. The storage space available is 10 m2 and a
box requires 0.5 m2 , a table takes up 1 m2 and a chair 0.4 m2 . Formulate this
problem as a linear programming problem for profit maximization and solve
it using simplex method.
Exercise 4. A carpenter makes boxes, tables and chairs. The profit contri-
SMA 2343
butions of the three products are $20, $30 and $10 respectively. The carpenter
can afford to spend up to 40 hours per week working and takes 2 hours to
make a box, 6 hours to make a table and 2 hours to make a chair. Customer
demand requires that he makes at most a third as many boxes as the total
number of chairs and tables. The storage space available is 10 m2 and a box
requires 1/2 m2 , a table takes up 1 m2 and a chair 2/5 m2 . Formulate this
problem as a linear programming problem for profit maximization and solve
it using simplex method.
Example. A certain company produces 3 products A, B C which contributes
a profit of Shs 8, 5 and 10 respectively. The production machine has 400 hrs.
Capacity and each product uses 2, 3 and 1 machine hour respectively. There
are 150 units available of a special component with A using 1 unit and C using
1 Unit per unit. A special allow of 200kg is needed in this period and product
A and C uses 2kg and 4kg per unit. There is a limitation of production of unit
B to not more than 50. Advice the company in order to maximize the profit.
Solution
Let x, y and z be the number of units of products A, B and C to be
produced respectively.
The model is:
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
24
Operations Research 1
Max p = 8x + 5y + 10z
2x + 3y + z 400 (Hours)
S/t
y 0,
z0
2x + 3y + z + s1 = 400
SMA 2343
x + z + s2 = 150
2x + 4z + s3 = 200
y + s4 = 50
x 0,
y 0,
z0
x3
s1
s2
s3
s4
Solution
S1
400
S2
150
S3
100
S4
50
10
Proceed and get the solution as x = 100, y = 50, x3 = 0 and Max p = 1050
3.7. Dealing with mixed on decision variables
If in a LP one or more constraint is given as xi k where k is a constant,
make the substitution yi = xi k 0 xi = yi + k.
We then replace xi with yi + k and adjust each constraint and the objective
function appropriately.
Example
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
25
Operations Research 1
M aximize Z = 5x1 + 3x2 + 4x3
Subject to 3x1 + 12x2 + 6x3 900
6x1 + 6x2 + 3x3 1350
2x1 + 3x2 + 3x3 390
x1 0, x2 20, x3 10
Since x2 and x3 cannot be zero, we let
a = x1 a 0
b = x2 20 b 0
a = x3 10 c 0
Adjusting the LP we get
All rights reserved
SMA 2343
M aximize Z = 5a + 3b + 4c + 100
Subject to 3a + 12b + 6c 900 240 60 = 600
6a + 6b + 3c 1350 120 30 = 1200
2a + 3b + 3c 390 60 30 300
a, b, c 0
which is now solvable using the simplex method. Following the simplex algorithm finally leads to the optimal table as
Basis a
s1
s2
s3
Solution
s1
5
2
1
2
21
50
s2
-1
-2
-1
100
1
2
5
2
150
3
2
7
2
3
2
9
2
850
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
26
Operations Research 1
Maximize Z = C 0 X subject to
AX b, X 0
If this is not, the case we need to transform the original problem to take
this standard form before we use ordinary simplex method
Example. Solve the LP below using simplex method
M inimize
z = 20x + 30y
subject to
x + y = 45
3x + 4y 170
12x + 6y 480
All rights reserved
SMA 2343
x 0, y 0
To write it in standard form,we
replace = with and inequalities
negate the objective function.
and solve it in the ordinary way as follows
Solution
Performing the adjustment we get
Maximize z = -20x -30y subject to
x + y <= 45
-x - y
<=-45
3x + 4y <= 170
12x +6y <= 480
x, y >=0
Adding slack variables we get
Maximize z = -20x -30y subject to
x + y + s1
= 45
-x - y +s2
= -45
3x + 4y + s3 = 170
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
27
Operations Research 1
12x +6y + s4 = 480
x, y >=0
The first tableau is given by
Tableau #1
Basis
s1
s2
s3
s4
Solution
s1
45
s2
-1
-1
-45
s3
170
s4
12
480
20
30
SMA 2343
Since all entries in the profit row are +ve, this solution is
optimal. But since s2=-45, the solution is infeasible
The pivot columns are those with -ve entries (column x and y)
to identify the entering variable take the ratios of the z-row
to the elements of the s2-row i.e (20/-1, 30/-1). The E.V is the
one corresponding to the smallest absolute value => x
Max x= min(+ve)(45/1, -45/-1, 170/3, 480/12)=40 from s4-row,
which implies that pivot value = 12
New row x=s4/12
Performing row reduction operations on x column gives
Tableau #2
JJ II
J
J DocDoc I
Back Close
Basis
s1
s2
s3
s4
s1
1/2
-1/12
s2
-1/2
1/12
-5
s3
5/2
-1/4
50
1/2
1/12
40
20
-5/3
-800
JKUAT
c M. Kihoro
J.
Solution
28
Operations Research 1
Since s2=-5, the solution is still infeasible
The pivot column is y and so y enters the solution
with pivot value -1/2
Max y=(-5/(-1/2)=10 (NOTE THAT THE OTHER RATIOS ARE NOT VALID)
New y-row=s2/(-1/2)
Reducing all the elements of y column to zero using Jordan
gauss computations gives
SMA 2343
Tableau #3
Basis
s1
s2
s3
s4
Solution
S1
-2
-1/6
10
s3
1/6
25
1/6
35
40
5/3
-1000
z = 2x + y
subject to
x+y =4
2x y 3
x 0, y 0
Optimal Solution: p =
JJ II
J
J DocDoc I
Back Close
JKUAT
; x=
, y=
c M. Kihoro
J.
29
Operations Research 1
Solutions to Examples
Exercise 3. The solution is as follows
Let x, y and z be the number of boxes, tables and chairs
to be produce respectively, then
Maximize profit p = 20x + 30y + 10z
subject to
SMA 2343
tabuleau leads to
Tableau #1
Basis x
s1
s2
s3
Soln
s1
40
s2
-1
-1
s3
1/2
2/5
10
-20
-30
-10
JJ II
J
J DocDoc I
Back Close
Basis x
s1
s2
s3
Soln
1/3
1/3
1/6
20/3
s2
10/3
-2/3
1/6
20/3
s3
1/6
1/15
-1/6
10/3
-10
200
JKUAT
c M. Kihoro
J.
30
Operations Research 1
Corresponding solution x=0, y=20/3, z=0, p=200 which is not
optimal. Entering variable corresponds to largest -ve value in
p-row =>x.
Maximum x=min((20/3)/(1/3),(20/3)/(10/3),(10/3)/(1/6))=2
corresponding to s2 row. Departing variable is s2,
pivot value =10/3
New row x=s2/(10/3)
Using the new row to reduce all other elements pivot column
to zero gives
SMA 2343
Tableau #3
Basis x
s1
s2
s3
Soln
2/5
3/20
-1/10
-1/5
1/20
3/10
s3
1/10
-7/40
-1/20
11/2
220
-2
JJ II
J
J DocDoc I
Back Close
Basis x
s1
s2
s3
Soln
5/2
3/8
-1/4
15
1/2
1/8
1/4
s3
-1/4
-17/80
-1/40
3/2
25/4
5/2
250
JKUAT
c M. Kihoro
J.
31
Operations Research 1
$250
Exercise 3
subject to
All rights reserved
SMA 2343
6y+
2z
3x + 1y+ 1z
1/2x +
40, (restrictiononhours)
0, (customersrestrictions)
1y+
x 0, y 0, z 0
We convert the problem to canonical form by introducing slack variables s1 ,
s2 and s3 to get
Maximize p = 20x + 30y + 10z
subject to
2x +
6y +
2z +s1 = 40,
6x + 1y +
1z +s2 =
0,
JJ II
J
J DocDoc I
Back Close
JKUAT
Basis
z s1 s2 s3 Soln
s1
40
s2
s3
1/2
2/5
10
p 20 30 10
c M. Kihoro
J.
32
Operations Research 1
This solution not optimal because there exists -ve values in the last row. Entering variable corresponds to the most negative value this row which in this
case is y
Maximum y = min( 40
, 0 , 10 ) = 0 which implies that pivot value = -1 and
6 1 1
departing variable is s2 , New row y = s2 / 1.
Performing row operations using the pivot row to reduce all the elements
of y column to zero leads to the following table.
Basis
x y
s1
20 0
40
3 1
7/2 0 (3)/5
10
0 30
SMA 2343
s3
z s1
p 110 0
20
s2 s3 Soln
This solution not optimal because there exists -ve values in the last row.
Entering variable corresponds to the most negative value this row which in
this case is x.
Maximum x = min( 40
, 0 , 10 ) = 2 which implies that pivot value = 20
20 3 7/2
and departing variable is s1 , New row x = s1 /20.
Performing row operations using the pivot row to reduce all the elements
of x column to zero leads to the following table.
Basis x y
s1
x 1 0 (1)/5
1/20
3/10
y 0 1
3/20 (1)/10
220
s3 0 0
p 0 0
2/5
2
s2 s3 Soln
11/2
This solution not optimal because there exists -ve values in the last row. Entering variable corresponds to the most negative value this row which in this
case is z
2
6
3
Maximum z = min( (1)/5
, 2/5
, 1/10
) = 15 which implies that pivot value =
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
33
Operations Research 1
Performing row operations using the pivot row to reduce all the elements
of z column to zero leads to the following table.
Basis x
y z
s1
x 1
1/2 0
1/8
1/4
z 0
5/2 1
3/8
(1)/4
15
3/2
250
p 0
5 0
25/4
s2 s3 Soln
5/2
Since all entries in the last row are negative, this table gives the optimal
solution as; x = 5, z = 15, s3 = 3/2 with a maximum objective function value
of p = 250.
All rights reserved
SMA 2343
JJ II
J
J DocDoc I
Back Close
JKUAT
c M. Kihoro
J.
Exercise 4
34