Topic 2 LINEAR PROGRAMING Notes

2.1 Definition of linear programming

It is a mathematical technique that deals with the optimization (maximization or minimization)
of a linear function of variables known as objective function to a set of linear inequalities
known as constraints.

2x+3y linear function 2x2+3y quadratic function 2x3+3y cubic function

It is a method of determining an optimum programme of inter-dependent activities in view of
available resources.

3x+4y≤18 linear enequalities

3x+4y=18 linear equation
2x+3y function

2.2 Basic parts of a linear programming problem

i. The objective function: it describes the primary purpose of the formulation i.e. to maximize
profit or to minimize cost.
ii. The constraint set: It is a system of equalities and / or inequalities describing the conditions
under which optimization is to be accomplished e.g. machine hours, man-hours, materials etc.
iii. It must have non negativity constrain- which show the solution cannot be zero for any of the
variable.

2.3 Assumptions of Linear programming

i. W e a s s u m e t h e r e l a t i o n s h i p b e t w e e n v a r i a b l e s s h o w s Linearity: Costs,
revenues or any physical properties, which form the basis of the problem, vary in direct
proportion with the quantities or numbers of components produced.
ii. Assumed the quantiles under consideration are Divisible: Quantities, revenues and costs
are infinitely divisible i.e. any fraction or decimal answer is valid.
iii. When using linear programming we assume a state Certainty : The technique makes no
allowance for uncertainty in the estimate made.
iv. Positive solutions: Non-negativity constraints are introduced to ensure only positive values
are considered.
v. Interdependence between demand products is ignored, products may be complementary or a
substitute for one another.
vi. Time factors are ignored. All production is assumed to be instantaneous.
vii. Costs and benefits which cannot be quantified easily, such as liquidity, good will and labour
stability are ignored.
2.4 Applications of Linear Programming
 Business and industry e.g. in the petroleum industry
 Food processing industry: to determine the optimal mix of feeds/ ingradients
 Paper and textile industry: To determine the optimal cutting method to minimize
trim losses
 Transport industry: To determine the best route
 Financial institutions – to determine investment plans.
19
Zetech Digital School
Operations Research
  Advertising media: assigning advertising expenditures to different media plans.
 Politics- campaign strategies
 Auditing – to find the number of financial audits
 Agriculture- amount of fertilizer to apply per acre
 Hospital scheduling- number of nurses to employ
 Marketing – determining
 Crude oil refining

2.5 Mathematical formulation of Linear Programming problems

If there are n decision variables and m constraints in the problem, the mathematical formulation
of the LP is:-

20
Zetech Digital School
Operations Research
 Optimization (Max or Min) Z c1 x1 + c2 x2 ......+ cn xn

Subject to the constraints

a11 x1 + a12 x2 +......... +a1n xn ≤ b1
a 21 x1 + a22 x2+ ............+ a2n xn≤ b2
am1 x1+ am2 x2+ ........+ a1n xn ≤ bm
and x1 , x2 ,........xn ≥ 0

Example
ABC chemical company is producing two products X and Y. the processing times are 3 hrs and 4
hrs per unit for X on operations one and two respectively and 4 hrs and 5 hrs per unit for Y on
operations one and two respectively. The available time is 18 hrs and 21 hrs for operations one
and two respectively. The product X can be sold at Sh. 3 profit per unit and Y at Sh.8 profit per
unit. Formulate the above as a LP problem.
Solution
Let x = one unit of product A and y = one unit of product B.
The constraints are the available times on machine one and two.
X Y Availability
Operation One 3 4 18
Operation two 4 5 21

Therefore the linear programming problem would be

Maximise Z=3X+8Y
Subject to :
3x+4y≤18
4x+5y≤21
x≤0, y≤
21
Zetech Digital School
Operations Research
Solutions to LP problems
LP problems can be solved by the help of the following methods:
i. Graphical method
ii. Simplex method
2.6 Graphical solution method
In order to solve a LP problem graphically, the following procedure is adopted:-
i. Formulate the appropriate LP problem
ii. Graph the constraint inequalities as follows: treat each inequality as though it were an
equality and for each equation, arbitrarily select two sets of points. Plot each of the points
and connect them with appropriate lines.
iii. Identify the solution space or feasible region which satisfies all the
constraints simultaneously. For ≤constraints, this region is below the lines and for ≥
constraints, the region is above the lines.
iv. Locate the solution points on the feasible region. These points always occur corner
points of the feasible region.
v. Evaluate the objective function at each of the corner points.
vi. Identify the optimum value of the objective function.
NB: It is assumed that the students covered the graphical solution method
Activity 2a
1. A company makes two products (1 and 2). Each product requires time on two machines A and
B. the specifications for each product are as follows

Product 1 Product 2
Processing time on machine A (hrs/unit) 2 2
Material and labour cost (Sh./ unit) 1 2
Material and labour cost (Sh./ unit) 14 15
Selling price (Sh./ unit) 16 18
Maximum possible sale (units) 130 150

The amount of time available on machine A is 360 hrs and on machine B is 260 hrs.

Formulate the above as an LP problem, and determine the number of units product 1 and 2,
which should be sold in order to maximize total contribution for the company.

2. ABC Company produces two products. Unit profit for product X is Sh. 60 and for product Y is
Sh. 50. Each product must pass through two machines P and Q. Product A requires 10 minutes
on machine P and 8 minutes on machine Q. Product Y requires 20 minutes on machine P and 5
minutes on Machine Q. machine P is available 200 minutes a day while machine Q is available
80 minutes a day. The company must produce at least two units of product X and five units of
product Yeach day. Units that are not completed in a given day are finished the following day.
22
Zetech Digital School
Operations Research
Required:
i. Formulate the above as a linear programming problem
ii. Solve using the graphical method and determine the most profitable daily production plan.

2.7 The Simplex method

The simplex method is an iterative algorithm for efficiently solving L.P problems. It was first
developed in 1947 by G.B. dantzig and his associates in the U.S. Department of the air force.

The simplex process

1. Standardize the problem into a linear programming tableau.
2. Generate the initial solution, called a basis.
3. Test the solution for optimality. If the solution is not optimal, improve it (go to step
4); otherwise go to step 6
4. Identify one variable that will leave the basis and one variable that will enter the basis.
5. Generate an improved solution. The improved solution is checked for optimality. If it
is not optimal, then steps 4 and 5 are repeated. If it is optimal, step 6 is undertaken.
6. Find if more than one optimal solution exists.

Standardizing the problem

Since only corner points of the feasible solution space are to be checked, and since these points
are defined by the intersection of equations, it is necessary to convert the inequalities in the
problem statement into equations in order to find the co-ordinates of the intersecting points. Such
a conversion depends on the type of constraints involved

Constraints with inequalities of the smaller- than- or-equal-to type

A slack variable is added to each of these type constraints.

23
Zetech Digital School
Operations Research
In the above example, there exists five unknown variables; x1, x2, s1, s2 and s3. and on the other
hand, there are only three equations. A solution to a system of linear equations requires that the
number of equations equal the number of variables, otherwise, there will be an infinite number of
solutions. To overcome this problem, two of the five variables are set to zero. Then it is possible
to solve the system of linear equations for the three remaining variables. In general, two cases are
distinguished:
1. Solutions with exactly m variables, where m is the number of constraints, are called basic
solutions. The m variables in these solutions are called basic variables and they constitute the
basis of the solution. The remaining variables have values of zero.
2. Solutions that include more than m nonzero variables are called nonbasic solutions. The basic
solutions can be further divided into two groups: feasible and infeasible. The feasible solutions
are those that satisfy all the constraints i.e. they are in the feasible area and the value of all their
variables must be nonnegative. The simplex algorithm searchesfor basic feasible solutions only.

A natural solution to this set equations would be to set x1 and x2 as the nonbasic variables with
s1, s2 and s3 as the basic variables. This is because x1 and x2 appear in more than one equation,
whereas s1, s2 and s3 each appear only once and in different equations and can thus be solved
easily. Setting x1 = x2 = 0 and solving s1, s2 and s3 ;
S1 =40, S2 =45 S3 =12
The resulting profit is found from the objective function as:-
Maximizie Z= 300x1+250x2+0s1+0s2+0s3

=300(0)+ 250(0+ 0(40)+0(45)+0(12)=0

23
Zetech Digital School
Operations Research
The Zj row: this row shows the amount of profits the objective function will be reduced by when
one unit of the variable, in each column is brought into the basis. Each value in the row is found
by: Multiply the elements in the “unit profits” column by the corresponding elements of the main
body, on a variable-by-variable basis and total the results. E.g. for column 2: 0(1) + 0(3) + 0(0) =
0.
The evaluator row: Cj – Zj. This row shows the net impact on the value of the objective
function of bringing one unit of each of the column variables into the basis. The Cj’s indicate
how much will be gained while the Zj’s indicate how much will be simultaneously lost. If the
difference between the two is positive, then the value of the objective function can be increased
by introducing one unit of the variable into the solution. The value of cj – zj represents the
opportunity cost of not having one unit of the corresponding variable in the solution.
Step 3: Test for optimality
Rule for optimality: For an optimal solution of a maximization problem, it is sufficient if all the
coefficients of the Cj – zj row are non-positive (either zero or negative). Since the solution is not
optimal, it can be improved by exchanging one variable in the solution. The improvement is
done by:
 Identifying the incoming variable
 Identifying the outgoing variable
 Building an improved tableau

Step 4: Identifying the incoming and the outgoing variable

The incoming variable: an incoming variable (currently nonbasic, to be changed to a basic
variable) is chosen by inspecting the Cj – Zj row of the current solution for that variable with the
largest positive coefficient. In our example X1 = 300. This is also known as the pivot column.

Rule for determining the outgoing variable:

Look down the column of the incoming variable and consider only positive elements
(coefficients). Then, divide the quantity of each constraint by the corresponding coefficient in the
column of the incoming variable. The row with the smallest ratio (it may be zero) is selected as
the outgoing row.
The coefficient of the incoming variable in the outgoing (pivot) row is called the pivot element.

24
Zetech Digital School
Operations Research
Step 5: Generating an improved solution
The solution is improved by introducing the incoming variable into the basis and removing the
outgoing variable. This is done row by row on the old the old tableau, forming, in the process, a
new one. The procedure starts with transforming the row of the outgoing variable, then the other
basis rows, and finally transforming the Cj, Zj and Cj- Zj rows.
Rule for Transformation of the outgoing row: Divide the main body elements and the ‘quantity”
of the outgoing row by the pivot element.
Transformation of the other rows
1. Identify the coefficient at the intersection of the row to be transformed and the incoming
column.
2. Multiply this number, in turn, by every element of the transformed outgoing row
3. Subtract the result of the second activity from the old row to derive the new transformed row.

Step 3: Test for optimality

The new solution must now be tested for optimality. It can be seen that the variable X2 in the
new Cj – Zj row has a positive coefficient, therefore the current solution is not optimal.
25
Zetech Digital School
Operations Research
Step 4(repeat): Identifying the incoming and the outgoing variable
X2 has the largest positive coefficient in the Cj – Zj row (250) and therefore it is the incoming
variable. S2 is the outgoing variable since it has the smallest nonnegative ratio.
Step 5: (repeat) Generate an improved solution
The rows of the tableau of the old solution are transformed one at a time to a new solution.

2.8 Advantages of Linear Programming

The advantages of linear programming are:-
i. Helps in attaining the optimum use of productive factors
ii. Improves the quality of decisions
iii. Improves the knowledge and skill of tomorrow’s executives
iv. It highlights the bottlenecks in the production process
v. It gives insight and perspective into problem situations
vi. Enables one to consider all possible solutions to problems.
vii. Enables one to come up with better and more successful decisions
viii. It’s a better tool for adjusting to meet changing conditions.

26
Zetech Digital School
Operations Research
2.9 Limitations of Linear Programming
The limitations of linear programming are:
i. It treats all relationships as linear
ii. It is assumed that any activity is infinitely divisible
iii. It takes into account single objective only i.e. profit maximization or cost
minimization
iv. It can be adopted only under the condition of certainty i.e. that the resources available,
per unit contribution, costs etc are known with certainty. This does not hold in real
situations.

LECTURE TWO EXERCISES: LINEAR PROGRAMMING

1. A firm makes two product X and Y and has a total production capacity of nine tonnes
per day, X and Y requiring the same production capacity. The firm has a permanent
contract to supply at least two tonnes of X and at least three tonnes of Y per day to
another company. Each tonne of X requires twenty machine hours production time
and each tonne of Y requires fifty machine hours production time. The daily
maximum possible number of machine hours is three hundred and sixty. The entire
firm’s output can be sold and the profit made is Sh. 80 per tonne of X and Sh. 120 per
tonne of Y.
Required:
i. Formulate the above as a linear programming problem.
ii. Using the graphical method, determine the number of tonnes of X and Y
that the company should produce, hence the maximum profit.
2. Reddy Mikks Company produces both interior and exterior paints from two raw
materials M1 and M2. The following table provides the basis data of the problem:-

Tonnes of raw material per tonne of Maximum daily

Exterior paint Interior Paint availability (Tonnes)
Raw Material M1 6 4 24
Raw Material M2 1 2 6
Profit per tonne 5 4
(Sh. 000)

A market survey indicates that the daily demand for interior paint cannot exceed that of
exterior paint by more than one tonne. Also the maximum daily demand of interior paint
is two tonnes. Reddy Mikks Company wants to determine the optimum (best) product
mix of interior and exterior paints that maximizes the total daily profit.
Required:-
i. Formulate the above as a linear programming problem.
ii. Using the graphical method, determine the optimum product mix hence the
optimum profit.

27
Zetech Digital School
Operations Research
3. A chicken farmer wishes to produce a chicken feed which is a blend of two inputs
coded P and Q. the table below shows the percentage of ingredients in each blend
input into the blend and the minimum daily requirements. Each unit of input P and Q
costs KSh. 40 and KSh. 10 respectively.
Proportion of input per unit of blend Minimum

Ingredients P Q requirements
Protein 30% 25% 400g
Carbohydrates 45% 15% 80g
Salt 15% 30% 120g
Water 5% 10% 200g

Required:
(a) Define the decision variables and the constraints for the above problem hence
formulate the problem as a linear programming model given the farmer’s
problem objective is to produce the blend that meets the minimum requirement
at least cost.
(b) Solve the above problem graphically hence indicate the optimal solution and the
optimal objective value.
4. An electronics firm is undecided as to the most profitable mix for its products. The
products now manufactured are transistors, resistors and electron tubes with a profit
(per 100 units) of Sh. 100, Sh. 60 and Sh. 40 respectively. To produce a shipment of
transistors containing 100 units requires one hour of engineering services, ten hours
of direct labour and two hours of administrative services. To produce 100 resistors are
required one hour, four hours and two hours of engineering, direct labour and
administrative time respectively. To produce one shipment of the electron tubes (100
units) requires one hour, five hours and six hours of engineering, direct labour and
administrative time respectively. There are 100 hours of engineering services, 600
hours of direct labour and 300 hours of administration available with the firm. What
is the most profitable mix?

5. A farmer has 1,000 acres on which he can grow corn, wheat or soya beans. Each acre
of corn costs Sh. 100,000 for preparation, requires 7 man-days of work and yields a
profit of Sh. 30,000. An acre of wheat costs Sh. 120,000 for preparation, requires 10
man-days of work and yields a profit of Sh. 40,000. An acre of soyabeans costs Sh.
70,000 for preparation, requires 8 man-days of work and yields a profit of Sh. 20,000.
If the farmer has Sh.100, 000,000 for preparation and can count on 8,000 man-day of
work, how many acres should be allocated to each crop to maximize profits?

6. An automobile company manufactures three models of cars. There is a back log of

orders with the company. Model A requires 60, 100 and 80 worker-days in three
production processes I, II and III respectively. Model B requires 100, 240 and 100
worker-days. Model C requires 200, 360 and 160 worker-days, respectively in three
28
Zetech Digital School
Operations Research
production processes. The numbers of workers employed in the three production
processes are 15, 30 and 15 respectively and an average worker is on the job for 200
working days a year. The expected profit for each model is Sh. 7500, 15000 and
30000 respectively. With this capacity, determine the company’s optimum product
mix and total profit.
 303
030
Zetech Digital School
Operations Research
 120
1201
20
Zetech Digital School