Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
History
In 1939, Leonid Kantorovich gave the first Linear Programming problem which is
identical to formulation of problem. The method to solve the same was given during the
World War II. It helped to minimize the losses, reduce the cost during the time and
maximize the loss to the enemy. Around the same time T.C. Koopmans, the Dutch
American economist gave the linear programming techniques and programs. They both
won the Nobel Prize.
Example 1
Arman requires 15, 17, and 19 units of pesticides A, B and C respectively for his farm. A
liquid pesticide contains 7, 4 and 6 units of A, B and C respectively per jar. A dry product
contains 2, 3 and 5 units of A, B and C per box.. If the liquid product sells for Rs.4 per
jar and the dry product sells for Rs.3 per carton, how many of each should be brought in
order to reduce the cost and meet the needs?
i) Give the decision variable of the problem?
Let X1 and X2 be the units of liquid pesticide and dry products required by Arman
Methods to solve
lpp
1 graphical solution method
Extreme point theorem forms the basis of this solution method
1 Extreme point theorem: It states that the extreme or the corner points of the
feasible area gives the most optimum and best output to the problem.
2 Feasible solution: The non negative set of values that satisfy the objective
function and all the constraints together is known as feasible solution.
3 Infeasible solution: These may or may not be negative values that do not satisfy
the constraints and objective function.
4 Basic solution: This is the solution obtained by equating (a-b) variables to zero
where b is the number equation.
5 Basic feasible solution: a basic solution which is also feasible.
6 Degenerate: When at least one basic variable is zero.
7 Non degenerate: when all basic variables are non-zero and positive.
8 Unbounded solution: A solution which may affect the objective function
indefinitely.
2)Simplex method
There are some cases when there are more than two decision variables associated with
the linear programming problems. When such things happen the solution to such
problems cannot be found using graphical solution method. The problem of linear
programming with two or more than three decision variables can be solved using a
systematic procedure called Simplex method. George B. Dantzig discovered the
solution algorithm for the simplex method. This method helps us to evaluate the corner
points such that each corner points gives us a better solution than the other one.
1 Standard form: When all the constraints are written in the form of equation then
the linear programming is known to be in the standard form
2 Slack variable: In order to transform the inequality into an equation a variable is
added to left hand side of a less than or equal side. This is known as slack
variable.
3 Surplus variable: In order to convert an inequality into an equation a variable is
subtracted from left hand side of a greater than or equal side. This variable is
known as surplus variable.
4 Basic solution: When solution is obtained by equating (a-b) variables equal to
zero where b is the number of equalities.
5 Simplex table: When simplex method is used a table is used to keep a track of
calculation made at each iteration.
6 Key column: The column containing largest positive number in Z J CJ row in
minimization case or largest negative number in ZJ CJ row in maximization case.
7 Key row: The row corresponding to smallest positive ratio
8 Key element: The intersection of key column and key row.
3)Big-m method
Artificial variables can be removed using the Big-G Method. It can be done by giving
coefficients to the artificial variables irrespective of the objective function. In case of
minimization a very large positive coefficient is assigned called as penalty, in the same
way in case of maximization a very large negative penalty is assigned.
Big-M Method
Steps to follow
1 Standard form is given to the problem.
2 Addition of non- negative variables. If there are no solution to the problem
these artificial variables will appear in the solution in the last even if we want
to eliminate them.
3 Initiation of table.
4 Optimization.
a ZJ - CJ 0, optimum basic feasible solution
b ZJ CJ 0, selection of key element and initiation of second table
5 Feasibility test
4)two phase method
In the first phase of this method, there is a minimization of the sum of all artificial
variables. In the second phase the objective function is optimized to give a feasible
solution. The method is known as two phase method since the method is completed
in two phase.
Algorithm steps
Phase 1- Transformation into Maximization type.
a Max. z 0: Existence of no feasible solution
b Max. z = 0 and presence of no artificial variable in the last row: Go for
phase 2 to obtain an optimal solution
c Max. z = 0 and presence of an artificial variable in the last row or the
basis: Begin with phase to get a feasible solution.
ij
ij
j=1
Thus, the total cost of transporting the commodity from all the sources to all
the destinations
is
m
Total cost =
cijxij
i=1
j=1
= c 11x11+c12x12+..+c1nx1n
c21x21+c22x22+..+c2nx2n
cm1xm1+cm2xm2+.+cmnxmn
In order to minimise the transportation costs, the following problem must be solved:
MINIMISE
z=
i=1
j=1
cijxij
SUBJECT TO
ij
j=1
ai for i=1n
AND
i=1
WHERE
ij
bj for j=1.n
The first constraint says that the sum of all shipments from a source cannot exceed
the available supply. The second constraint specifies that the sum of all shipments
to a destination must be at least as large as the demand. The above implies that
m
i=1
b
j=1
When the total supply is equal to the total demand then the transportation model is
said to be balanced. In a balanced transportation model, each of the constraints is
an equation:
n
x
j=1
ij
= ai for i=1.m
x
i=1
ij
= bj for j=1.n
A transportation model in which the total supply and total demand are unequal is
called unbalanced. It is always possible to balance an unbalanced transportation
problem.
Example 1- Balanced transportation model. Consider the following problem with 2
factories and 3 warehouses:
Factory 1
Factory 2
Demand
Warehouse 1
C11
C21
7
Warehouse 2
C12
C22
10
Warehouse 3
C13
C23
13
Supply
20
10
1. Suppose the demand at warehouse 1 above is 9 units. Then the total supply and
total demand are unequal, and the problem is unbalanced. In this case it is not
possible to satisfy all the demand at each destination simultaneously. We
reformulate the model as follows: since demand exceeds supply by 2 units, we
introduce a dummy source (i.e. a fictitious factory) which has a capacity of 2. The
amount shipped from this dummy source to a destination represents the shortage
quantity at that destination. It is necessary to specify the costs associated with the
dummy source. There are two situations to consider
(a) Since the source does not exist, no shipping from the source will occur, so the
unit transportation costs can be set to zero.
(b) Alternatively, if a penalty cost, P, is incurred for every
unit of
unsatisfied demand, then the unit transportation costs
should be set equal to
the unit penalty costs.
FACTORY 1
FACTORY 2
DUMMY
DEMAND
WAREHOUSE
1
C11
C21
P
7
WAREHOUSE
2
C12
C22
P
10
WAREHOUSE
3
C13
C23
P
13
SUPPLY
20
10
2
FACTORY 1
FACTORY 2
DEMAND
WAREHOUS
E1
C11
C21
7
WAREHOUS
E2
C12
C22
10
WAREHOUS
E3
C13
C31
13
DUMMY
SUPPLY
S
S
4
20
10
Here we are allocating the excess supply to the different destinations. From now on,
we will discuss balanced transportation problems only, as any unbalanced problem
can always be balanced according to the above constructions
a
i=1
b
j=1
W1
W2
W3
W4
supply
F1
10
20
11
20
F2
12
20
25
F3
14
16
18
25
demand
10
15
15
20
2. If a column (or row) is satisfied, cross it out. The remaining decision variables in
that column (or row) are non-basic and are set equal to zero. If a row and column
are satisfied simultaneously, cross only one out (it does not matter which).
3. Adjust supply and demand for the non-crossed out rows and columns.
4. Allocate the maximum feasible amount to the first available non-crossed out
element in the next column (or row).
5. When exactly one row or column is left, all the remaining variables are basic and
are assigned the only feasible allocation
. For the above example:
x11 = 10. Cross out column 1. The amount left in row 1 is 10.
x12 = 10. Cross out row 1. 5 units are left in column 2.
x22 = 5. Cross out column 2. 20 units are left in row 2.
x23 = 15. Cross out column 3. 5 units are left in row 2.
Only column 4 is now left and so both the remaining variables x 24 and x34 will be
basic. The only feasible allocation of the 5 units in row 2 and the 15 units in row 3 is
to allocate x24 = 5 and x34 = 15, which also ensures that the demand in column 4 is
satisfied. 5 This provides the initial basic feasible solution x 11 = 10, x12 = 10, x22 = 5,
x23 = 15, x24 = 5, x34 = 15. The remaining variables are non-basic and therefore
equal to zero. The solution has m + n 1 = 6 basic variables as required. The
values of the basic variables xij are entered in the top left of each cell. There should
always be m + n 1 of these; in certain (degenerate) cases some of them may be
zero. They must always add up to the total supply and demand in each row and
column. Note that some books position the data differently in the cells of the
tableau.
F1
F2
F3
Demand
W1
10
12
0
10
W2
0
7
14
15
W3
20
9
16
15
W4
11
20
18
20
supply
20
25
15
shadow price). If any sij is negative (so that i+ j > cij ), then the total cost can be
reduced by allocating as many units as possible to cell (i, j). However, if all the s ij
are positive then it will be more expensive to change any of the allocations and so
we have found a minimum cost.
Thus the procedure is as follows:
1. Assign values of i and j to the columns.
2. Enter the values sij = cij i j in every cell.
3. If all the sij s are non-negative, we have an optimal solution.
Assigning values of i and j to our example with the initial basic feasible solution
given by the North-West Corner method, gives the following transportation tableau:
Adding the sij variables to each cell, we find three negative values and so the
solution is not optimal.
Iterating the algorithm
If the current solution is not optimal, we need a method for moving to a better
basic feasible solution. As previously, this involves changing only one variable in the
basis so again we must identify an entering variable and a departing variable in the
basis
Determining the entering variable
If the current solution is not optimal, choose the cell with the most negative value
of sij as the entering variable, as the cost will be reduced most by using this route.
For our example, the most negative value is s31 and so the entering variable is x31.
Determining the leaving variable
We construct a closed loop that starts and ends at the entering variable and
comprises successive horizontal and vertical segments whose end points must be
basic variables (except those associated with the entering variable). It does not
matter whether the loop is clockwise or anticlockwise.
Starting tableau
We now see how large the entering variable can be made without violating the
feasibility conditions. Suppose x31 increases from zero to some level > 0. Then x 11
must change to 10 to preserve the demand constraint in column 1. This has a
knock on effect for x12 which must change to 10 + . This process continues for all
the corners of the loop. The departing variable is chosen from among the corners of
the loop which decrease when the entering variable increases above zero level. It is
the one with the smallest current value, as this will be the first to reach zero as the
entering variable increases. Any further increase in the entering variable past this
value leads to infeasibility. Clearly x22 is the departing variable in this case. The
entering variable x31 can increase to 5 and feasibility will be preserved. New values
of the i s and the j s can now be assigned and the test for optimality applied. If
the solution is still not optimal, new entering and departing variables must be
determined and the process repeated.
Second tableau
Third tableau
We construct i s, j s and sij s as before, and then check for optimality. The
tableau is not optimal as x22 is negative and is therefore the entering variable. The
loop construction shows that can be as large as 10, and that x 24 is the departing
variable.
Fourth tableau
Final Solution
The minimum cost is given by 5 0 + 15 11 + 10 7 + 15 9 + 10 0 + 5
18 = 460 which occurs when x12 = 5, x14 = 15, x22 = 10, x23 = 15, x31 = 10, x34= 5
and all the other decision variables are equal to zero. 7.2.4 Solving the
transportation problem with Excel Solver We can use Excel Solver to solve the
transportation problem. We set the problem out in the general form of a linear
programming problem
The data is entered as two arrays - one of transportation costs and the other as
decision variables. The Excel spreadhseet illustrating this is available on the module
website.
We can use linear programming in numerous fields of our life and not only that we
use techniques of linear programming in many fields of our life be it from our house
budget to the government policies, economy, and business etcetera.
Important areas of application of linear programming are:
1 Manufacturing problems: It helps us to find the optimum methods in order to
maximize the profit, to know about amount of raw material, machinery, labour
resource needed.
2 Assembling problem: To combine things in order to have a feasible and optimum
solution.
3 Diet problem: Linear programming helps us to determine the exact and optimum
constituents and nutrients needed by the body to have a healthy life.
4 Blending problem:
5 Transportation problems: It helps us to determine the schedule and way in order
to have cheapest mode of transportation.
6 Investment problems: Linear programming helps us to invest in number of fixed
income to help us make maximum profit.
Linear programming can be used to express many practical problems. Network flow
problem and multicommodity flow problem are some special cases of linear
programming. Duality, decomposition, convexity are some concepts of optimization
theory that are central to linear programming. Linear programming is used in
company management, microeconomics, technology, planning, production, and so
more.
Linear programming characterizes so many issues be it from our daily life basis or to
the scientific basis.
Conclusion
In this project our main motto is to discuss the basic concepts about Linear
Programming and its Application. Also by using the concepts of Linear Programming
we discuss its application in real life. This tells us that the technique of linear
programming can be used to obtain optimal solution. We could apply it in our real
lives to have good and feasible solutions.