21

22

23


24


25

26

27

28


29
REDDY MIKKS
x1: #tons exterior paint x2: #tons interior paint daily
starting tableau:
basic z x1 x2 s1 s2 s3 s4 RHS solution ratios row operations select
z 1 -5 -4 0 0 0 0 0 0

6r1+5r2
most negative
s1 0 6 4 1 0 0 0 24 24 4

minimum ratio 0
s2 0 1 2 0 1 0 0 6 6 6 6r3-r2

s3 0 -1 1 0 0 1 0 1 1 -1 6r4+r2
ignore ratio <0
s4 0 0 1 0 0 0 1 2 2

ignore ratio


basic z x1 x2 s1 s2 s3 s4 RHS solution ratios row operations select
z 6 0 -4 5 0 0 0 120 20

2r1+r3
most negative
x1 0 6 4 1 0 0 0 24 4 6 2r2-r3

s2 0 0 8 -1 6 0 0 12 2 1.5

minimum ratio 0
s3 0 0 10 1 0 6 0 30 5 3 4r4-5r3

s4 0 0 1 0 0 0 1 2 2 2 8r5-r3



basic z x1 x2 s1 s2 s3 s4 RHS solution ratios row operations select
z 12 0 0 9 6 0 0 252 21

OPTIMAL
x1 0 12 0 3 -6 0 0 36 3


x2 0 0 8 -1 6 0 0 12 1


s3 0 0 0 9 -30 24 0 60 2


s4 0 0 0 1 -6 0 8 4




basic z x1 x2 s1 s2 s3 s4 RHS solution


z 1 0 0 0 0 21 21

shadow prices

x1 0 1 0 - 0 0 3 3

*$1000 scarce
x2 0 0 1 - 0 0 1 1

*$1000 scarce
s3 0 0 0 -1 1 0 2 2

0 abundant
s4 0 0 0 - 0 1

0 abundant

Simplex method
Eight types of variables:
Decision variables are to be solved with the LP-model.
Slacks, surplus, artificial variables turn the LP-model into equation form.
Basic variables are positive variables in some iteration.
Nonbasics are zero variables in some iteration.
Leaving variable exits the list of basics becoming nonbasic and equal to zero.
Entering variable enters list of basics becoming positive.

All constraints must have RHS 0
All variables, including slacks, surplus and artificial variables, must be 0

Add slacks and surplus variables to each row, including the z-row, so that each row is an equation.
Turn the z-row in an equation with RHS =0
Now set up the starting tableau. The left-most column contains the basics in some iteration.

30
Iterations:
Max problem (Min problem): Select the most negative (positive) entry of the z-row: pivot column
Select minimum ratio >0, ignoring zero or negative ratios: pivot row. Now you found the pivot!

The pivot column gives the entering variable, the pivot row gives the leaving variable.

Conduct Gauss-Jordan row operations turning all entries in the pivot column into zeros (except for
the pivot itself) as follows:
Let a be the pivot and b is the entry in row j which should be made equal to zero.
Change row j to a*row j b*pivot row.
In this way you never use fractions, which are indeed annoying and a source of mistakes!
Now update the column of basics: the leaving variable is replaced with the entering variable

Repeat these steps until the z-row has no negative entries for a maximum problem or no positive
entries for a minimum problem (except for degenerate problems).

Basics in the left-most column also belong to columns which have all entries but one equal to zero.

The shadow price of a means of production (a constraint) is equal to
Az
A capacIty
, in $/kg or some
other unit. The shadow price is the extra profit per extra unit capacity.
Shadow prices can be found in the z-row as column headings of the slacks or surplus variables.

Shadow price =0 means that there is no extra profit for extra capacity. This can only be the case if
there is capacity left; the slack is positive, so a basic. This means of production is called abundant.

Shadow price >0 means that there is extra profit for extra capacity. This is the case if the capacity
is fully consumed; the slack is zero, so a nonbasic. This means of production is called scarce.

Shadow prices are relevant under the condition that the optimal production combination does not
change (to another corner point).

Inspect last tableau of Reddy Mikks:
The value of each basic (z included) in each iteration can be found as follows:
RHS
pot
where the piv-
ot is the entry at the intersection of the row and column of that variable.
All other variables are nonbasics and equal to zero.

Decision variables x1, x2
Basic: x1, x2, s3, s4: columns with but one entry 0 or look in the le-most column.
Nonbasic: s1, s2: columns with more than one entry 0 or not menoned in the le-most column.

Because s1 =0, capacity of M1 is fully consumed. More profit is only possible if the capacity of M1
is raised, giving $750 extra profit per extra ton M1. This is the shadow price, which can be found in
the heading of column s1 in the z-row. M1 is a scarce means of production.
If from M2 1 ton more is available, profit increases with $500/ton M2, shadow price is in the head-
ing of column s2. M2 is scarce as well.
Because s3 =2.5 and s4 =0.5, these two slacks are positive, meaning that there is capacity left, so
that market and demand limits have not been fully consumed yet, market and demand are called
abundant. Their shadow price is zero: there will be no extra profit if we raise these limits.
31
The headings of columns s3 and s4, which contain shadow prices of market and demand are zero.
Final tableau once more:

32



3.4 ARTIFICIAL STARTING SOLUTION
As demonstrated, LPs with all constraints and with non-negative right-hand sides offer a con-
venient all-slack starting basic feasible solution. Models involving =and / or constraints do not.
The procedure for starting ill-behaved LPs with =and / or constraints is to use artificial varia-
bles that play the role of slacks at the first iteration, and dispose of them at a later iteration.
A relevant and widely used method to achieve this is the two-phase method.

3.4.2 TWO PHASE METHOD
As the name suggests, the method solves the LP in two phases: Phase I attempts to find a starting
basic feasible solution, and, if one is found, Phase II is invoked to solve the original problem, start-
ing with the result of Phase I.

Summary of the Two-Phase Method
Phase I: Put the problem in equation form, add the necessary artificial variables for = and / or
constraints. Whether the LP is maximization or not, ALWAYSuse as objective MINIMIZATION
OF THE SUM OF THE ARTIFICIALS: r =R1 +R2 +......
If the minimum value of this sum is positive, the LP-model has no feasible solution.
Otherwise, proceed to Phase II.

Phase II: Use the feasible solution of Phase I, replace the r-row with the original z-row and remove
the columns of all artificial variables as a starting tableau for phase II.
33












Basic r x1 x2 x3 R1 R2 x4 RHS Solution ratio operations
r 1 0 0 0 -1 -1 0 0 0 r1+r2
R1 0 3 1 0 1 0 0 3 3 3
R2 0 4 3 -1 0 1 0 6 6
x4 0 1 2 0 0 0 1 4 4

r 1 3 1 0 0 -1 0 3 r1+r3
R1 0 3 1 0 1 0 0 3 3
R2 0 4 3 -1 0 1 0 6 6 6
x4 0 1 2 0 0 0 1 4 4
34
Phase I: basics are artificials R1 =3 and R2 =6 and slack x4 =4. The first two iterations show row
operations to change the entries in the r-row of the columns of the basics (artificial) in zeros.
Because the objective is to minimize r =R1 +R2, select the most positive entry in the r-row. In two
iterations phase I is optimized with feasible solution r =0.
Now all artificials are nonbasics and have to be removed from the tableau.
Replace the r-row with the original z-row.
The solution of the original problem is: x1 =0.6, x2 =1.2, x3 =0. It can be checked that this is the
first feasible corner solution, so phase I brings the solution of the LP-model in the feasible region.

Phase II: Once more turn entries in the z-row of the columns of the basics x1, x2 and x4 to zeros.


































After these row operations, phase II is optimized in one iteration, starting in the feasible solution
where phase I left us. Because the original problem is minimizing, look for the most positive entry
in the z-row. Entering variable is x3, leaving variable is x4.
Remark: Practically all commercial packages use the Two Phase Method to solve the LP-models.

r 1 7 4 -1 0 0 0 9 9 ratio 3r1-7r2
R1 0 3 1 0 1 0 0 3 3 1
R2 0 4 3 -1 0 1 0 6 6 1.5 3r3-4r2
x4 0 1 2 0 0 0 1 4 4 4 3r4-r2
r 3 0 5 -3 -7 0 0 6 2 ratio r1-r3
x1 0 3 1 0 1 0 0 3 1 3 5r2-r3
R2 0 0 5 -3 -4 3 0 6 2 1.2
x4 0 0 5 0 -1 0 3 9 3 1.8 r4-r3
r 3 0 0 0 -3 -3 0 0 0 r is optimal
x1 0 15 0 3 9 -3 0 9 0.6 r2/3
x2 0 0 5 -3 -4 3 0 6 1.2
x4 0 0 0 3 3 -3 3 3 1 r4/3

Basic z x1 x2 x3 x4 RHS Solution operations
z 1 -4 -1 0 0 0 0

5r1+4r2
x1 0 5 0 1 0 3 0.6


x2 0 0 5 -3 0 6 1.2


x4 0 0 0 1 1 1 1


z 5 0 -5 4 0 12 2.4 r1+r3
x1 0 5 0 1 0 3 0.6


x2 0 0 5 -3 0 6 1.2


x4 0 0 0 1 1 1 1


Basic z x1 x2 x3 x4 RHS Solution ratio operations
z 5 0 0 1 0 18 3.6

r1-r4
x1 0 5 0 1 0 3 0.6

3 r2-r4
x2 0 0 5 -3 0 6 1.2

ignore r3+3r4
x4 0 0 0 1 1 1 1

1
z 5 0 0 0 -1 17 3.4 optimal
x1 0 5 0 0 -1 2 0.4


x2 0 0 5 0 3 9 1.8


x3 0 0 0 1 1 1 1
35



36



37

38




















39







40



41


42

43



44
45




























46

47

48








49








50


51


52






53
3.6.4 Sensitivity Analysis with Solver
We now have all the tools needed to decipher the output provided by the LP software, par-
ticularly with regard to sensitivity analysis. We will use the TOYCO example to demonstrate
the Solver output.

54
the screen. You can then click Sensitivity Report I to view the results.
We will use the Excel output for sensitivity analysis. The calculations on preceding pages about
sensitivity analysis show the meaning of the numbers in Sensitivity Report.




55
Problems Chapter 3
Fig 2.1 Feasible space of the Reddy Mikks model Fig 2.2 Optimum solution of the Reddy Mikks model.

1 In the Reddy Mikks model (ex. 2.2-1), consider the feasible solution x
1
(ext.paint) =3 tons and x
2

(int. paint) =1 ton. Determine the value of the slacks for raw materials M1 and M2.

2






a In the diet model (ex 2.2-2), determine the
surplus amount of feed consisting of 500 lb
of corn and 600 lb of soybean meal.
b Solve the diet model with the simplex
method. Start with the tableau of phase I
and optimize phase I. Then optimize with
phase II. Compare your solution of course
with the solution in the graph.

3 A company produces two products A and B. The sales volume for A has to be at least 80% of the
total sales of both A and B. However, the company cannot sell more than 100 units of A per
day. Both products use one raw material, of which the maximum daily availability is 240 kg. The
usage rates of the raw material are 2 kg p.u. of A and 4 kg p.u. of B. The profit per unit for A and
B is 20 and 50, respectively. Determine the optimum product mix for the company that
wants to maximize the profit. Solution: 2600; 80, 20

4 An assembly line consisting of three consecutive stations
produces two home cinema sets HCS1 and HCS2. The adja-
cent table provides the assembly times for the three work-
stations. The daily maintenance for stations 1, 2 and 3 takes
up, respectively, 10%, 14% and 12% of the maximum 480
minutes, available per station per day.
Determine the optimum product mix that minimizes the idle times in the three work stations.
minutes per unit
workstation HCS1 HCS2
1 6 4
2 5 5
3 4 6
Hinimizc z =0.3x
1
+0.9x
2

sub]cct to
x
1
+ x
2
800
0.21x
1
0.30x
2
0
0.03x
1
0.01x
2
0

x
1
,x
2
0
56
5 A company produces textile rolls with a standard width of 120 cm, where each roll has a fixed
length. A customer orders 50 rolls of 40 cm wide and 60 rolls of 25 cm. These rolls have to be
cut from rolls of standard width 120 cm. The firm may decide to cut a standard roll in a roll of
40 cm wide and three rolls of 25 cm. This results in a trim loss of 5 cm. The standard rolls have
to be cut in such a way that a minimum trim loss is guaranteed. Superfluous rolls are also con-
sidered as trim loss (see problem 7 Ch1_2). Minimize the trim loss with the simplex method.

6 Solve the following problems with the simplex method:

a
Hox z = 3x
1
+x
2
3x
3
+5x
4
x
1
+2x
2
+2x
3
+4x
4
40
2x
1
x
2
+x
3
+2x
4
10
4x
1
2x
2
+x
3
x
4
10
b
Hox z = 8x
1
+6x
2
+9x
3
2x
4
x
1
+2x
2
+2x
3
+4x
4
40
2x
1
x
2
+x
3
+2x
4
10
4x
1
2x
2
+x
3
x
4
10


c
Hox z = 5x
1
x
2
+3x
3
+4x
4
x
1
+2x
2
+2x
3
+4x
4
40
2x
1
x
2
+x
3
+2x
4
10
4x
1
2x
2
+x
3
x
4
10
d
Hox z = 5x
1
4x
2
+6x
3
8x
4
x
1
+2x
2
+2x
3
+4x
4
40
2x
1
x
2
+x
3
+2x
4
10
4x
1
2x
2
+x
3
x
4
10


7 The Gutchi company manufactures purs-
es, shaving bags and backpacks. The con-
struction includes leather and synthetics,
leather being a scarce raw material. The
production process requires two types of
skilled labour: sewing and finishing. The
following table gives the availability of the resources, their usage by the three products, and the
profits per unit. Solve with the simplex method, check with the solver of Excel.

8 Solve the following LP models with the Two-Phase Method.

a
Hox z = x
1
+3x
2
x
1
+2x
2
3
x
1
+x
2
1
x
1
+x
2
5
b
Hin z = x
1
+3x
2
x
1
+2x
2
3
x
1
+x
2
1
x
1
+x
2
5


9 Wild West produces two types of cowboy hats. A type 1 hat requires twice as much labour time
as a type 2 hat. If all the available labour time is dedicated to type 2 alone, the company can
produce a total of 400 type 2 hats a day. The respective market limits for the two types are 150
and 200 hats a day.
The revenue is $8 per type 1 hat and $5 per type 2 hat.
a Solve graphically the maximum revenue and the optimal number of hats per type.
b Determine the shadow price / dual price of the production capacity in terms of type 2 hat
and the range for which it is applicable.
c If the daily limit on the type 1 hat is decreased to 120, use the shadow price to determine
the corresponding effect on the optimal revenue.
d What is the dual price of the market share of the type 2 hat? By how much can the market
share be increased while yielding the computed worth per unit?
e Determine the optimality range for the unit revenue ratio of the two types of hats that will
keep the current optimum unchanged.
f Will the optimal solution change if the revenue per unit is the same for both types?

resource requirements per unit
resource purse bag backpack daily availability
leather (ft
2
) 2 1 3 42
sewing (hr) 2 1 2 40
finishing (hr) 3 1 2 45
selling price $ 24 22 45
57
10 In the Reddy Mikks model Example 2.2-1, x1 =tons ext.paint, x2 =tons int.paint.
a Determine the range for the ratios of the unit revenues of exterior paint and interior paint.
b (1) The revenue per ton exterior paint becomes $4000/ton. Determine the optimality range
of the unit revenue of interior paint that will keep the optimal solution unchanged.
(2) Answer the same question for exterior paint if the revenue per ton interior paint remains
constant at $3500/ton
c If for marketing reasons the unit revenue of interior paint must be reduced to $3000, will the
current optimum production mix change?

11 Consider the TOYCO model (3.6-2, pg 44).
a Suppose that any additional time for operation 1 beyond its current capacity of 430 minutes
per day must be done on an overtime basis at $50 an hour. The hourly cost includes both la-
bour and the operation of the machine. Is it economically advantageous to use overtime
with operation 1?
b Suppose that the operator of operation 2 has agreed to work 2 hours of overtime daily at
$45 an hour. Additionally, the cost of the operation itself is $10 an hour.
What is the net effect of this activity on the daily revenue?
c Is overtime needed for operation 3?
d Suppose that the daily availability of operation 1 is increased to 440 minutes. Any overtime
used beyond the current maximum capacity will cost $40 an hour.
Determine the new optimum solution and net revenue.
e Suppose that the daily availability of operation 2 is decreased by 15 minutes a day and that
the hourly cost of the operation during regular time is $30.
Is it advantageous to decrease the availability of operation 2?

12 HiDec produces two models of electronic gadg-
ets that use resistors, capacitors and chips. The
following table summarizes the data of the situ-
ation. The amounts produced of model 1 and 2
are x1 and x2, respectively.
See attached optimal tableau to the right.
a Produce the LP model.
b Determine each resource status
(scarce / abundant).
c Determine the shadow price of each resource.
d Determine the feasibility range of each shadow
price.
e The available number of resistors increases to 1300. Find the new optimum solution.
f The available number of chips decreases to 350. Is it possible to determine the new opti-
mum solution directly from the given information?
g The available number of capacitors is limited to the range in (d). Determine the correspond-
ing range of the optimal revenue and ranges for the units to be produced of model 1 and 2.
h A new contractor is offering to sell HiDec additional resistors at 40 cents each, but only if Hi-
Dec would purchase at least 500 units. Should HiDec accept the offer?

unit resource requirements (units)
resource model 1 model 2 max. availability
resistor 2 3 1200
capacitor 2 1 1000
chips 0 4 800
unit price $ 3 4
basic x1 x2 s1 s2 s3 solution
z 0 0 1.25 0.25 0 1750
x1 1 0 -0.25 0.75 0 450
s3 0 0 -2 2 1 400
x2 0 1 0.5 -0.5 0 100