e-Notes by P.V.

Raveendra, MSRIT, Bangalore

O5 MBA 21 QUANTITATIVE TECHNIQUES FOR MANAGEMENT MODULE 1 INTRODUCTION TO OPERATIONS REASERACH

Origin of Operations Research was during the Second world war. The military management in England called upon a team ( inter disciplinary ) of scientists to study the strategic and tactical problems related to air and land defense of the country. Since they were having very limited military resources, it was necessary to decide upon the most effective utilization of them e.g., the efficient ocean transport, effective bombing. The OR teams were not actually engaged in military operation and in fighting the war. But, they were only advisors and significantly instrumental in winning the war to the extent that the scientific and systematic approaches involved in OR provided a good intellectual support to the strategic initiates of the military commands. Hence OR can be associated an art of winning the war without actually fighting it. The OR teams in U.S. are known as Operational analysis, operations evaluation, system analysis operations research and management science. Following the end of war, the success of military teams attracted the attention of industrial manager who were seeking solution to their complex executive type problems. The most common problem was: what methods should be adopted so that the total cost is minimum or total profits maximum? Today, the impact of OR can be felt in man areas apart from military and business application. Which include transportation system, libraries, hospitals, city planning, financial institutions etc. OR provides a quantitative technique or a scientific approach to the executives for making better decisions for operation under their control. In other words the OR provides a scientific approach to problem solving for executive management. While making use of the techniques of OR, a mathematical model of the problem is formulated. This model is actually a simplified representation of the 1

problems in which only the most important features are considered for reasons of simplicity. Then an optional or most favorable solution is found. Since the model is an idealized instead, the solutions requires further improvement. The apparent weaknesses in the initial solutions are used to suggest improvement in the model, its input data and the solution procedure. A new solution is thus obtained and the process is repeated until the further improvement in the succeeding solutions becomes negligible. If the model is carefully formulated and tested, the resulting solution should reach to be good approximation to the ideal course of action for the real problems. Although, we may not get the best answers, but definitely we are able to find the bad answers where worse exist. The OR techniques are always able to save us from worse situations of practical life. Definitions: OR is the systematic method oriented study of the basic structure, characteristics, functions and relationships of an organization to provide the executive with a sound, scientific and quantitative basis for decision making OR is the application of scientific methods, techniques and tools to problems involving the operations of system so as to provide these in control of the operations with optimum solutions to the problems OR is concerned with scientifically deciding how to best design and operate man-machine systems usually requiring the allocation of scarce resources.

Characteristics of OR: OR approaches problem solving and decision making from a total systems perspective. It is interdisciplinary. Model building and

mathematical manipulation provide the methodology which has been the key contribution of OR. OR is for operations economy. The primary focus is on decision making and computers are used

PHASES OF OPERATIONS RESEARCH STUDY

PHASE 1; Formulation The Problem:

Formulate the problem in the form of appropriate model by using the following information 1. Who has to take the decision? 2. What are the objectives? 3. What are the ranges of controlled variables? 4. What are the uncontrolled variable that may affect the possible solutions? 5. What are the restrictions or constraints on the variable?

PHASE II Constructing a Mathematical method: Reformulate the problem in an appropriate form with suitable identification of both static and dynamic structural elements. A mathematical model should include the following three important basic factors. i) ii) iii) Decision variable and parameters Constraints or restrictions Objective function.

PHASE III : Deriving the solutions from the model: Compute those values of decision variable that maximize /minimize the objective function. Such solution is called optimal solution which is always in the best interest of the problem under consideration. The general techniques for deriving the solution of OR model includes Distribution models ( LPP, transportation and assignment), Queuing models ,Network analysis, Job sequencing , Replacement models, Simulation models and etc.

PHASEIV: Testing the model and solution ( updating the model)

After completing the model, it is once again tested as a whole for the errors if any. A valid and good model can provide a reliable prediction of the systems performance and applicable for present and future specifications of the problem.

PHASEV: Controlling the solution: This phase establishes controls over the solution with any degree of satisfaction. The model requires immediate modification as the controlled

variables change significantly, otherwise the model goes out of control. As the conditions are constantly changing in the world, the model and the solution may not remain valid for a long time.

PHASE VI: Implementing the solution: Finally, the tested results of the model are implemented to work. This phase is primarily executed with cooperation of OR experts and those who are responsible for managing and operating the systems.

PRINCIPLE OF MODELLING 1. Do not build up a complicated model when simple one will suffice 2. Beware of molding the problem to fit the techniques 3. The deduction phase of modeling must be conducted rigorously 4. models should be validated prior to implementation 5. A model should never be taken too literally 6. A model should neither be pressed to do, nor criticized for failing to do that for which it was never intended 7. Beware of over selling a model 8. Some of the primary benefits of modeling are associated with process of developing the model 9. A model can not be any better than the information that goes into it 10. Models cannot replace decision makers.

Techniques of OR
1. Distribution ( allocation) models: Distribution models are concerned with the allotment of available resources so as to minimize cost or maximize profit subject to prescribed restrictions. Methods for solving

such type of problems are known as mathematical programming techniques. Which can be distinguished as linear a(LPP) and non-liner programming models(transportation and Assignment models) 2. Waiting Line( queuing) models: In using models an attempt is made to predict i) ii) iii) How much average time will be spent by the customer in a queue What will be an average length of waiting line or queue? What will be the traffic intensity of a queuing system? Etc

This models provides us methods to minimize the some of the cost of providing service and cost of obtaining service which are primarily associated with value of time spent by the customer in a queue.

3. Production/Inventory models. These are concerned with determination of optimal( economic order quantity and ordering interval considering the factors such a s demand per annum, cost of placing orders, inventory carrying cost. Such models also helpful in dealing with quantity discounts and multiple products. 4. Competitive strategy models( Games theory) these models are used to determine the behavior of decision making under competition or conflict. Methods for solving such models have not been found suitable for industrial application, mainly because they are referred to idealistic world neglecting any essential features of reality. 5. Network analysis. These models are applicable in large projects involving complexities and interdependencies of activates. PERT and CPM are used for planning scheduling and controlling complex project which can be characterized by networks.

6. Job Sequencing model: These models involve the selection is such a sequence of performing a series of jobs to be done on service facilities(machines) that optimize the efficiency measure of performance of the system. 7. Replacement models: these deal with determinations of optimum replacement policy in situation that arise when some items of machinery need replacement by a new one. Individual and group replacement

policies can be used in the case of such equipment that fail completely and instantaneously. 8. Markovian models: These models are applicable in such situating where the state of the system can defined by some description measure of numerical value and where the system moves from one state to another on a probability basis. Brand switching problems considered in marketing studies ins an example of this model. 9. Simulation models: these models are more useful when following problem arises. 1) the number of variables and constraints relationship may be so large that it is not computationally feasible to pursue such analysis. 2. this model may be much away from the reality that no confidence can be placed on the computational results. In fact such

models are solved by simulation techniques where no other methods are available for its solution.

Scope of Operations Research:

OR is useful in the following various important fields In Agriculture: With explosion of population and consequent shortage of food every country is facing the problem of a) Optimum allocation of land to various corps in accordance with e climatic conditions b) Optimum distribution of water from various resources like can for irrigation purposes. In Finance: OR techniques can be applied i) to maximize the per capita income with minimum resources

ii) iii)

to find out the profit plan for the company to determine the best replacement policies

In Industry: OR is applicable in deciding optimum allocation of various limited resources such as men, machine , material ,money, time to arrive at the optimal decision. In Marketing: OR can be applied i) Where to distribute the products for sale so that the total cost of transportation is minimum ii) iii) iv) v) The minimum per unit sale rice The size of the stock to met the future demand How to select the best advertising media with respect to time, cost etc. How, when and what to purchase at the minimum possible cost.

5. In Personnel management: OR can be applied i) To appoint the most suitable persons on minimum salary ii) to determine the best age of retirement for the employees iii)To find out the number of persons to be appointed on full time basis when the workload is seasonal. In Production Management: A production manager can use OR techniques i) ii) To find out number and size of the items to produced In scheduling and sequencing the production run by proper allocation of machines iii) iv) In calculating the optimum product mix To select, locate and design the sites for production plants.

Finally we can say: wherever there is a problem there is OR. The application of OR covers the whole extern of any thing. LIMITATIONS: Magnitude of computation : Absence of Quantification: Distance between managers and OR experts.

REPLACEMENT THEORY PROBEMS

Problem 1:

The cost of a machine is Rs 6100 & its scrap value is only Rs

100. Maintenance costs are followed from the experience.

Year Maintenance cost in Rs

1 100

2 250

3 400

4 600

5 900

8 2000

1250 1600

When should machine be replaced?

Solution:
Total Difference cost- Total cost

year

Maintenance cost 100 250 400 600 900 1250 1600 2000

Maintenance B/N cost 100 350 750 1350 2250 3500 5100 7100

Average cost 6100 3175 2250 1837.5 1650 1583.33 1585.7 1637.5

scarp value 6000 6000 6000 6000 6000 6000 6000 6000 6100 6350 6750 7350 8250 9500 11100 12100

1 2 3 4 5 6 7 8

As the average cost is lowest in the 6th year machine should replaced at the end of 6th year or at the beginning of the 7th year.

Problem 2: A machine owner finds from his past records that cost per year
of maintaining a machine whose purchase price is Rs 6000 are given below,

Year Maintenance cost in Rs Resale price

5 2300 200

8 4000 200

1000 1200 1400 1800 3000 1500 750 375

2800 3400 200 200

Determine at what each is a replacement due.

Solution:
Total Difference cost- Total cost

year

Maintenance cost 1000 1200 1400 1800 2300 2800 3400 4000

Maintenance B/N cost 1000 2200 3600 5400 7700 10500 13900 17900

Average cost 4000 3350 2950 2756 2700 2717 2814 2962.5

scarp value 3000 4500 5250 5625 5800 5800 5800 5800 4000 6700 8850 11025 13500 16300 19700 23700

1 2 3 4 5 6 7 8

As the average cost is less in the 5th year machine should replaced at the end of the 5th year or at the starting of the 6th year.

Problem 3.

The cost of a new machine is Rs.5000. The maintenance cost

during the nth year is given by Rn = 500 X (n-1), n=1,2,Suppose that the

discount rate per year is 0.05 . After how many years it will be economical to replace the machine by a new one?

Solution:
Since the discount rate of money per year is 0.05, the present worth of the money to be spent over in a period of one year is v = 1/(1 + o.o5) = 0.9523

For the best replacement age, we compute the following table for the given machine by using present worth or one rupee to be spent in n year from now onwards Computation of weighted average cost for machine

Year

Maintenan ce Cost Rn

Discount Factor V=(1/1+r)

n

Discounted Maitenanc e Cost 4= (2) X (3)

Discounted Cumulati cumulative Cost C+RnV

1 n-

Weighted average annual cost (6)/(5)

ve Discount ed Cost V n-1

(2)

(3)

(5)

(6)

1 2 3 4 5 6

0 500 1000 1500 2000 2500

1.0000 0.9523 0.9070 0.8638 0.8227 0.7835

0 476 907 1296 1645 1959

5000 5476 6383 7679 9324 11283

1.000 1.9523 2.8593 3.7231 4.5458 5.3293

5000 2805 2232 2063 2051 2117

10

From the above table, we find tht the minimum value of column ( 7) occurs after 5 years, hence it will be economical to replace the machne at the end of the 5 th year.

Problem 4:

The management of a large hotel is considering the periodic

replacement of light bulbs fitted in ins rooms. There are 500 rooms in the hotel and each room has 6 bulbs. The management is now following the policy of replacing the bulbs as they fail at a total cost of Rs.3 per bulb. The management feels that this cost can be reduced to Rs.1 by adopting the periodic replacement method. On the basis of the information give below, evaluate the alternative and make a recommendation to the management. Months of use 1 2 3 4 5

% of bulbs failing by that month

10

25

50

80

100

Solution :
Let Pi be the probability that a light bulb, which was new when installed fails This value is the difference between the during the ith month of its life.

proportion alive at the beginning of ith month and proportion alive at the end of ith month. Thus we have P1 = 10/100= 0.10 P2 = 25 10 /100 =0.15 P3 = (50 25)/100 =0.25 P4 = (80 50)/100 = 0.30 P5 =(100 80)/100 = 0.20 ( since the sum of al probabilities can never be greater than unity, all the probabilities beyond P5 should be taken as zero)

Let No denotes the number of original bulbs and Fi denotes the number of bulbs replacement made at the end of ith month. Then we have 11

No = 500 X 6 = 3000 F1 = No X P1 = 3000 X 0.1 = 300 i.e. the number of bulbs being replaced by the end of the first month. F2 = No. of bulbs replaced by the end of second month = ( number of bulbs in the beginning )( probability that these bulbs in the second month) + ( number of bulbs replaced in the first month) ( probability that these fails during second month) = No X P2 + F1 X P1 = ( 3000 X 0.15) + ( 300 X 0.1) = 480

F3 =No X P3 + F1 X P2 + F2 XP1 = ( 3000 X 0.25) + ( 300 X 0.15) + (480 X 0.10) = 843 F4 = ( 3000 X 0.3) + ( 300 X 0.25) + (480 X 0.15) + (843 X 0.10) = 1131 similarly F5 = 1049 Calculation of Average cost per month in group replacement policy

End the month

of Total number Bulbs failed

Cumulativ e Number of failures

Cost of Replacem ent ( Rs.3/bulb)

Cost of Group

Total Cost

Averag e cost per month

Replaceme (in Rs.) nt ( Re.1?bulb)

1 2 3 4 5

300 480 843 1131 1049

300 780 1623 2754 3803

900 2340 4869 8262 11409

3000 3000 3000 3000 3000

3900 5340 7869 11262 14409

3900 2670 2623 2815.5 2881.8

12

The average cost of is minimum in third month.

Individual replacement policy

Expected life of each bulb = i X Pi = 1 X 0.1 + 2 X 0.15 + 3 X 0.25 + 4 X 0.3 + 5 X 0.2 = 3.35 MONTHS. The average number of replacement will be = N / mean life = 3000 /3.35 = 896bulbs individual replacement = 896 X 3 = 2688. Decision: Thus it is better to have a individual replacement policy for first two months and group replacement after end of third month.

Problem 5: Following mortality rates have been observed for a certain type of
fuses Week 1 2 3 4 5

%of failure at the end of 5 week

15

35

75

100

There are 1,000 fuses in use and it costs Rs.5 to replace an individual fuse. If all fuses were replaced simultaneously if would cost Rs.1.25 per fuse. It is

proposed to replace all fuses at fixed intervals of time, whether or not they have burnt out, and to continue replacing burnt-out fuses as they fail. At what

intervals, the group replacement should be made? Also prove that this optimal policy is superior to the straight forward policy of replacing each fuse only when it fails.

13

Solution

:Let

Pi

be the probability that a fuse, which was new when

installed fails during the ith week of its life. This value is the difference between the proportion alive at the beginning of ith week and proportion alive at the end of ith week. Thus we have

P1 = 5/100= 0.05 P2 = 15 5 /100 =0.10 P3 = (35 15)/100 =0.20 P4 = (75 35)/100 = 0.40 P5 =(100 75)/100 = 0.25 ( since the sum of al probabilities can never be greater than unity, all the probabilities beyond P5 should be taken as zero)

Let No denotes the number of original fuses and Fi denotes the number of fuses replacement made at the end of ith week. Then we have No = 1000 F1 = No X P1 = 1000 X 0.05 = 50 i.e. the number of fuses being replaced by the end of the first week. F2 = No. of fuses replaced by the end of second week = ( number of fuses in the beginning )( probability that these fuses in the second week) + ( number of fuses replaced in the first week) ( probability that these fails during second week) = No X P2 + F1 X P1 = ( 1000 X 0.1) + ( 50 X 0.1) = 102

F3 =No X P3 + F1 X P2 + F2 XP1 = ( 1000 X 0.20) + ( 50 X 0.10) + (102 X 0.05) = 210 F4 = 210 similarly F5 = 430

14

Calculation of Average cost per week in group replacement policy

Cumulati End of Total ve the number week fuses failed Number of failures 1 50 50 2 3 4 5 102 210 430 333 152 362 792 1125

Cost of Replaceme nt ( Rs.5/fuse) 250 760 1810 3960 5625

Cost of Group Replacemen t ( re.1/fuse) 1250 1250 1250 1250 1250

Total Cost (in Rs.)

Averag e cost per week 1500 1005 1020 1302.5 1375

1500 2010 3060 5210 6875

The average cost of is minimum in second week. Individual replacement policy

Expected life of each bulb = i X Pi = 1 X 0.05 + 2 X 0.10 + 3 X 0.20 + 4 X 0.40 + 5 X 0.25 = 3.7weeks. The average number of replacement will be = N / mean life = 1000 /3.7 = 270 fuses Individual replacement = 270 X 5 = 1350. Decision: Thus it is better to have an individual replacement policy for first week and group replacement after end of second week.

15

Problem 6: A computer contains 10,000 resistors. When any of the register

fails it is replaced. The cost of replacing a single resistor is Rs.10 only. If all the resistors are replaced at the same time, the cost per resistor would be reduced to Rs.3.50. The per cent surviving by the end of the month t is as follows. What is the optimum plan. month %surviving at the end of month 0 100 1 97 2 90 3 70 4 30 5 15 6 0

Here % of survival rate is given with the help of it we can find the failure rate as follows. Failure in first month 100-97/100 =3

Failure in second month 100 90 /100 = 10 Similarly in the third month =30 Failure in the fourth month =70 Failure in the 5th month =85 Failure in the six month = 100

With the help of above failure rate one can find probabilities and continue in the similar manner( as above two problems).

16

Notes for Assignment models Dr. P.V.Raveendra Professor, M.S.R.I.T. SOLUTION TO THE PROBLEM 1

Programmers

120

100

80

90

80

90

110

70

110

140

120

100

90

90

80

90

1. CHECK THE OBJECTIVE FUNCTION ( GIVEN PROBLEMS IS MINIMISATION) 2. CHECK FOR THE BALANCE ( GIVEN PROBLEM IS BALANCED ONE) 3. NOW APPLY HUNGARIAN METHOD Row conversion matrix

A 1 2 3 120-80=40 80-70=10 110-100=10

B 20 20 40

C 0 40 20

D 10 0 0

17

90-80=10

10

10

COLUMN CONVERSION MATRIX

A 1 2 3 4 40-10=30 0 0 0

B 20-10=10 10 30 0

C 0 40 20 0

D 10 0 0 10

ALLOCATION ACCORDING TO THE METHOD

A 1 2 3 4 40-10=30

B 20-10=10 10 30

D 10 0x

0
40 20 0x

0
0x 0x

0
10

Allocation 1-

C 2-A 3-D 4 B

80 + 80 + 100 + 90 = 350

SOLUTION TO THE PROBLEM 2

1. CHECK THE OBJECTIVE FUNCTION ( GIVEN PROBLEMS IS MINIMISATION) 18

2. CHECK FOR THE BALANCE ( GIVEN PROBLEM IS BALANCED ONE) 3. NOW APPLY HUNGARIAN METHOD

II

III

IV

44

80

52

60

B C D

60 36 52

56 60 76

40 48 36

72 48 40

ROW REDUCTION I II III IV

36

16

B C D

20 0 16

16 24 40

0 12 0

32 12 4

COLUMN REDUCTION I A 0 II 20 III 8 IV 12

20

28

19

C D

0 16

8 24

12 0

8 0

ALLOCATION ACCORDING TO HUNGARIAN METHOD

I A B C D

II 20

III 8 0 12

IV 12 28 8 0X

0
20 0X 16

0
8 24

Mark the unallocated rows ( i.e. C) and mark the columns where there are zeros in the marked rows. (i.e.). Mark the rows where allocation is there in the marked columns (i.e. A). Now draw the lines through unmarked rows i.e.B and D and marked columns (i.e.) Selected the minimum element(8) among uncovered elements ( where lines are not passing) and deduct it from the rest of the uncovered elements .Add the same where ever there is intersection of the lines i.e. cell (2,1) and cell (4,1) The revised table is as follows with allocations I A B C II 12 III 0 X 0X 4 IV 4 28

0
20+8=28 0X

0
0X

20

16+8=24

24

0X

THUS A I , B II, CIV, D III WITH COST OF ( 44+56+48+36)

SOLUTION TO THE PROBLEM 3

1. CHECK THE OBJECTIVE FUNCTION ( GIVEN PROBLEMS IS MINIMISATION) 2. CHECK FOR THE BALANCE ( GIVEN PROBLEM IS BALANCED ONE) 3. NOW APPLY HUNGARIAN METHOD

M1 301 302 303 304 305

M2 4 1

M3 2 5 1 3 4

M4

M5 1 2

1 2 3

1 4 3 2

2 3

ROW REDUCTION

M1 301 302 303 304

M2 3 0

M3 1 4 0 1

M4

M5 0 1

0 1 1

0 3 1

21

305

COLUMN REDUCTION

(IT WILL BE THE SAME ABOVE AS THERE IS A ZERO IS EACH COLUMN)

M1 301 302 303 304 305

M2 3 0

M3 1 4 0 1 2

M4

M5 0 1

0 1 1

0 3 1 0

0 1

NOW ALLOCATING ACCORDING TO HUNGARIAN METHOD

M1 301 302 303 304 305

M2 3 0X

M3 1 4

M4

M5

0X 3 1

0
1

0
1 1

0
1 2

0
1

ALLOCATION 301 - M5 =1

22

302 M1 = 1 303- M3 =1 304 - M2 = 2 305- M4 =2

SOLUTION TO PROBLEM 4
THIS IS PROBLEM IS MAXIMASATION CASE Salesmen D1 D2 D3 D4 D5

S1 S2 S3 S4 S5

40 48 49 30 37

46 32 35 46 41

48 36 41 49 48

36 29 38 44 43

48 44 45 44 47

CONCVERTING INTO MINIMIZATION CASE BY SELECTION THE MAXIMUM NUMBER(49) FROM THE MATRIX AND DEDUCTING IT FROM THE CELL ELEMENTS ( LOSS MATRIX) WILL BE AS FOLLOWING ONE

Salesmen

D1

D2

D3

D4

D5

S1 S2

49-40=9 1

3 17

1 13

13 20

1 5

23

S3 S4 S5

0 19 12

14 3 8

8 0 1

11 5 6

4 5 2

2. CHECK FOR THE BALANCE ( THE ABOVE ONE IS BALANCED ONE AS IT IS SQUARE MATRIX) AS THIS PROBLEM IS MINIMISATION CASE WE CAN APPLY HUNGARIAN METHOD ROW CONVERSION Salesmen D1 D2 D3 D4 D5

S1 S2 S3 S4 S5

8 0 0 19 11

2 16 14 3 7

0 12 8 0 0

12 19 11 5 5

0 4 4 5 2

COLUMN REDUCTION Salesmen D1 D2 D3 D4 D5

S1 S2

8 0

0 14

0 12

7 14

0 4

24

S3 S4 S5

0 19 11

12 1 5

8 0 0

6 0 0

4 5 2

ALLOCATION

Salesmen

D1

D2

D3

D4

D5

S1 S2 S3

0
14 12

0X 12 8

7 14 6

0X 4 4

0
0X

S4 S5

19 11

1 5

0
0X

0X

5 2

Put a tick mark for unallocated rows( i.e. S4 ) AND in the marked row mark the column where there is zero ( i.e. D1) . In the marked column put a tick mark to the row where there is an allocation ( i.e. S2). Then draw the line through unmarked rows (i.e.S!,S4 and S5) and marked columns ( i.e.D1) Select the smallest element from the uncovered elements( where lines are not passing ) i.e 4 and deduct it from the rest of uncovered elements only. Add 4 where ever there is intersection of the lines ( i.e. cell (1,1) and (1,4) and (1,5)) Rewrite the table as follows and allocate according to the method.

25

Salesmen

D1

D2

D3

D4

D5

S1 S2 S3

8+4=12

0
14-4=10 12-4=8

0X 12-4=8 8-4=4

7 14-4=10 6-4=2

0X 0X

0
0X

0
5 2

S4 S5

19+3=23 11+4=15

1 5

0
0X

0X

NOW WE HAVE MADE ALLOCATION IN THE LOSS MATRIX. TAKE THESE ALLOCATION TO MAXIMISATION MATRIX AS SHOWN BELOW

Salesmen

D1

D2

D3

D4

D5

S1 S2 S3 S4

40

46
32 35 46

48 36 41

36 29 38 44

48 44

48
49 30

45
44

49

26

S5

37

41

48

43

47

THUS S1 TO D2 ,S2-D1,S3-D5,S4-D3 AND S5-D4 ( 46 +48+45+49+43) SALES.

SOLUTION TO PROBLEM 5

ASSUMING A TO A DISTANCE IS and so on.

From/to

A B C D E

4 7 3 4

7 6

3 3 7

4 4 5 7

6 3 4

7 5

Given problem is minimization case and balanced one. We can apply Hungarian method Row conversion matrix From/to A B C D E

A B

4 3

0 0

1 1

27

C D E

2 0 0

1 0 0

4 1

0 4

Column conversion matrix

From/to

A B C D E

1 2 0 0

3 2

0 0 2

1 1 0 4

1 0 0

3 0

Allocation matrix

From/to

A B C

1 2

3 2

0
0X 2

1 1 0

28

D E

0
0X

0X 0X

As explained in the above cases the improved matrix will be

From/to

A B C D E

0X 2

1 0X 0X

2 1

0X

0X 0X

0
3

0
4

0
0X

IN THE ABOVE SOLUTION D IS ASSIGNED TO A WHICH MEANS SALESMAN IS COMING BACK TO A WITHOUT VISITING ALL THE CITIES.. But the condition is that salesman should not come to home with out visiting to all the cities ( this condition is given priority compared to the cost , so reallocating as below the cost may increase by one rupee but the first condition is satisfied)

From/to

A B

0X

0
0 X

0X 0X

29

C D E

2 0

3 0X

0
4

0
0X

HERE ALLOCATION IS MADE TO CELL (2,3) WHERE COST IS 1 BECASUSE SALEMAN SHOULD NOT RETURN HOME WITHOUT VISITOING ALL THE CITIES. TOTAL COST IS A D-BCEA Rs. (3+3+6+5+4)

SOLUTION TO PROBLEM 6
We have to calculate the time required to complete the job P by A = 199/12 =17 hours. For each hour he will charge 5 rupees. So the cost involved in completion of the job P by A is given by 17 x 5 = 85.

typist A

R 65

S 85

T 75

199/12=17 75 X5=85

B C D E

90 75 80 70

78 66 72 64

66 57 60 56

90 75 80 72

78 69 72 68

The above problem is minimization case and balanced one. We can apply Hungarian method and find the solution as usual

30

SOLUTION TO PROBLEM 7
Let us assume that all the crew is based at Hyderbad. Then lay-over times for various combinations of flights as shown below. For example the flight A1 which start from Hyderabad at 6 a m , reached Delhi at 8 a m. If it is to return as B1, schedule time for which is 8 a m, they it can do so after 24 hours, since a minimum layover time of 5 hours is required. Similarly, lay over times for other flight combinations can be obtained. Layover time -- crew based at Hyderabad

Flight A1 A2 A3 A4

B1 24 22 16 10

B2 25 23 17 11

B3 6 28 22 16

B4 11 9 27 21

The lay-over times for various flight combinations, when crew is assumed to be based at Delhi, are similarly calculated as shown below

Lay-over time crew based at Delhi

Flight A1 A2 A3 A4

B1 20 22 28 10

B2 19 21 27 9

B3 14 16 22 28

B4 9 11 17 23

31

Now, since the crew can based at either of the stations, minimum lay-over times can be obtained for different flight combinations by selecting the corresponding lower value out of the above two tables. For instance in combining A1-B1, we select 20, which is lower of the two values 24 and 20.thus rewriting the table.

Layover times crew based at Hyderabad/Delhi

Flight A1 A2 A3 A4

B1 20* 22 16 10

B2 19* 21* 17 9*

B3 6 16* 22 16

B4 9* 9 17* 21

when crew is based at Delhi.

Now we can apply the Hungarian method to minimize the layover time and the allocation will be as follows Flight FLIGHT NO. LAYOVER TIME A1 A2 A3 A4 B3 B4 B1 B2 6 9 16 9 CREW IS BASED AT HYDERABAD HYDERABAD HYDERABAD DELHI.

** ************************************************************* Note: IF THE GIVEN PROBLEM IS UNBALANCED i.e. the number of row is not equal to the number of columns then we can add the column or rows depending on the requirement using a dummy row/column using zero as the cost. Then apply the Hungarian method. 32

Queuing Models Queuing System: General Structure Arrival Process According to source According to numbers According to time Service System Single server facility Multiple, parallel facilities with single queue Multiple, parallel facilities with multiple queues Service facilities in a parallel Queue Structure First come first served Last come first served Service in random order Priority service Model 1: Poisson-exponential single server model infinite population Assumptions: Arrivals are Poisson with a mean arrival rate of, say Service time is exponential, rate being Source population is infinite Customer service on first come first served basis Single service station For the system to be workable, Model 2: Poisson-exponential single server model finite population Has same assumptions as model 1, except that population is finite
33

 Model 3: Poisson-exponential multiple server model infinite population Assumptions Arrival of customers follows Poisson law, mean rate Service time has exponential distribution, mean service rate There are K service stations A single waiting line is formed Source population is infinite Service on a first-come-first-served basis Arrival rate is smaller than combined service rate of all service facilities Model: 1Operating Characteristics a) Queue length average number of customers in queue waiting to get service b) System length average number of customers in the system c) Waiting time in queue average waiting time of a customer to get service d) Total time in system average time a customer spends in the system e) Server idle time relative frequency with which system is idle Measurement parameters = mean number of arrivals per time period (eg. Per hour) = mean number of customers served per time period Probability of system being busy/traffic intensity = / Average waiting time system Ws = 1/(- ) Average waiting time in queue
34

Wq= / (- ) Average number of customers in the system Ls = / (- ) Average number of customers in the queue Lq = 2/ (- ) Probability of an empty facility/system being idle P(0) = 1 P(w) Probability of being in the system longer than time (t) P(T>t)= e (- )t Probability of customers not exceeding k in the system P (n.k) = k P( n>k) = (k+1) Probability of exactly N customers in the system P(N) = N (1-)

Problem 1. Customers arrive at a booking office window, being manned by a single individual rate of 25per hour. Time required to serve a customer has exponential distribution with a mean of 120 seconds. Find the mean waiting time of a customer in the queue Solution: Given =25 customer/hr = 30 customer/hr

Mean waiting time in Queue = / (- ) = 25/30 (30-25) = 0.167 hr Mean waiting time in system =1/ - =1/30-25 =0.2 hr Problem 2: A repairman is to be hired to repair machines which breakdown at a n average rate of 6 per hour. The breakdowns follow Poisson distribution. The non-production
35

time of a machine is considered to cost Rs. 20 per hour. Two repairmen Mr. X and Mr.Y have been interviewed for this purpose. Mr. X charges Rs.10 per hour and he service breakdown machines at the rate of 8 per hour. Mr. Y demands Rs.14 per hour and he services at an average of 12 per hour. Which repairman should be hired? ( Assume 8 hours shift per day) solution: Given =6 /hr x=8/hr y=12/hr Given no of machine cost at idle = 20Rs/hr No of machine in X Lsx = / x - =6/8-6 =3 machines Total no of machines =3*8=24 machines Total cost=hiring charges of x + cost of idle machine =10*8 + 24*20 =Rs.560 No of machine in y Lsy = / y =6/12-6 =1 Total no of machine = 1*8=8 machines Total cost= hiring charges of y + cost of idle machine = 14*8 + 20*8 = Rs.272 We chose Mr.Y since cost is lower than Mr. X Problem 3: A warehouse has only one loading dock manned by a three person crew. Trucks arrive at the loading dock at an average rate of 4 trucks per hour and the arrival rate is Poisson distributed. The loading of a truck takes 10 minutes on an average and can be assumed to be exponentially
36

distributed . The operating cost of a truck is Rs.20 per hour and the members of the crew are paid @ Rs.6 each per hour. Would you advise the truck owner to add another crew of three persons?

Solutions:

1) =4 truck/ hr =10min=60/10=6truck/hr

No of trucks in system, Ls= / (- ) =4/6-4 =2 trucks Total cost = cost of maintaining trucks + cost of crew =20*20 + 3*6 =40 + 18 = Rs.58 2) =4trucks /hr If we double the crew, = 12 trucks/ hr No of trucks in system, Ls= / (- ) =4/12-4 =0.5 trucks Total cost = cost of maintaining trucks + cost of crew =0.5*20 + 6*6 =10 + 36 =Rs.46 We will advice the owner to add 3 persons for loading cost the total cost for 3 extra persons is less than previous. Problem 4 At a service counter of fast-food joint, the customers arrive at the average interval of six minutes whereas the counter clerk takes on an average 5 minutes for preparation of bill and delivery of the item. Calculate the following
37

a. counter utilization level b. average waiting time of th4e customers at the fast food joint c. Expected average waiting time in the line d. Average number of customers in the service counter area e. average number of customer in the line f. probability that the counter clerk is idle g. Probability of finding the clerk busy h. chances that customer is required to wait more than 30 minutes in the system i. probability of having four customer in the system J) probability of finding more than 3 customer in the system

Solutions: Given = 60/10 = 10 customer/hr = 12 customer/hr a) = / = 10/12 = 0.833 b) Ws = 1/ - = 1/12-10 = 0.5 hr c) Wq = / (- ) = 10/12(12-10) = 0.416 hr d) Ls= / (- ) = 10/12-10 = 5 customers e) Lq = 2 / (- ) = 102 /12(12-10) = 4.167 customers f) 1- = 1- / = 1- 10/12 = 0.167 g) = / = 10/12 = 0.833 h) chances of probability that customer wait more than 30min = 30/60 = 0.5 hrs P (T>t) = e (- ) t
38

P (T>0.5) = e (12- 10) 0.5 = 0.368 i) P (N) = N (1-) P (4) = 4 (1-) = (0.833)4(1-0.833) = 0.0806 P (n>k) = (k+1) P (n>3) = (3+1) = ( / ) 4= (10/12) 4 = 0.474

j)

Problem 5: Customers arrive at a one-window drive-in bank according to a Poisson distribution with mean 10 per hour. Service time per customer is exponential with mean 5 minutes. The space in front of the window including that for the serviced car accommodate a maximum of 3 cars. Other cars can wait outside the space. Calculate A) what is the probability that an arriving customer can drive directly to the space in front of the window. B) what is the probability that an arriving customer will have to wait outside the indicated space C) How long is arriving customer expected to wait before stating the service. D) How many spaces should be provided in front of the window so that all the arriving customers can wait in front of the window at least 20% of the time.

solution: Given = 10 /hr = 5min = 60/5 = 12customer/hr = / = 10/12 = 0.83 We know that P (N) = N (1-) a) P = Po + P1 + P2 = 0 (1-) + 1 (1-) + 2 (1-) = (1- ) (1+ + 2) = (1- 0.83) (1+ 0.83 + 0.832)
39

= 0.428 b) P = 1 (Po + P1 + P2 + P3 ) = 1 [0.42 + 3 (1-)] = 1 [0.42 + 0.833 (1-0.83)] = 0.482

c) Wq = / (- ) = 10/12(12-10) = 0.416 hr d) Po = 0 (1-) 1-10/12 = 0.16 e) P1 = 1 (1-) = 10/12(1-10/12) = 0.14

Probability that there will be no or one car in the space which cover = Po + P1 = 0.30 = 30%, hence there should at least 1 car space for waiting at the window space at least 20% of the time.

Problem 6 Customers arrive at the first class ticket counter of a theatre at a rate of 12 per hours. There is one clerk serving the customers at a rate of 30 per hour. Assuming the conditions for use of the single channel queuing model, evaluate a) The probability that there is no customer at the counter (i.e. that the system is idle) b) The probability that there are more than 20 customers at the counter c) The probability that there is no customer waiting to be served d) The probability that a customer is being served and no body is waiting.

40

Solution: Given = 12 customers /hr = 30 customers/hr 1) Ideal =1- = 1- / = 1- 12/30 = 0.6 2) At least 3 customers at counter P (n>k) = (k+1) = (3+1) = (12/30)4 = 0.025

3) Probability that no customers waiting to be served P (at least 1 customer at the counter) = Po + P1 = 0 (1-) + 1 (1-) = (1- ) (1 + ) = (1- 2 ) = (1- (12/30)2 = 0.84 4) Probability that a customers being served and nobody waiting = P1 = 1 (1-) = 12/30(1-12/30) = 0.24 Give feedback to raveendrapv@rediffmail.com

41

42