Download as pdf or txt
Download as pdf or txt
You are on page 1of 21

QUA N T T A E MODULE

Simulation

DISCUSSION QUESTIONS 7. Monte Carlo simulation is a technique that uses random


numbers to develop values for random variables described by
1. The seven steps of simulation are: define the problem;
appropriate probabi lity distributions:
introduce the important variables associated with the problem;
The steps in developing a Monte Carlo simulation are:
construct a numerical model; set up possible courscs of action for
tcsting; run the experi ment; consider the resu Its ; and decide what • Step I: Establish a probability distribution for each random
course of action to take. variable
• Step 2: Build a cumulative probability distribution for each
2. Advantages of simulation:
random variable
I. It is relatively straightforward and tlexiblc. • Step 3: Establish an interval of random numbers for each
2. It can be used to analyze large and complex real-world random variable
situations that cannot be solved by conventional operations • Step 4: Generate random nllmbers
management models. • Step 5: Simulate a series of trials
3. It allows for the inclusion of real-world complications that
8. Simulation can be used in bllsilless in dozens of ways, from
most models cannot permit.
examining lines in banks and post offices, to testing inventory
4. It allows "time compression" .
policies. to layout of a plant. to scheduling employees and parts.
5. It allows the user to ask "what if" questions and
etc. Table F.I provided some applications.
experiment witb various representations of the problem.
6. It does not interfere with the real world system. 9. Simulatioll is widely IIsed because complex real-world sys­
7. It allows us to study interactive effects of individual tems can be examined and tested without impacting on the actual
components or variables. situation. It also allows for "time-compression." allows "what-if"
analysis. and is more and more easily modeled with a new
3. Disadvantages of simulation:
generation of simulation packages.
I. It can be very expensive and complex.
10. Special-purpose languages have these advantages:
2. It does not generate optimal solutions to problems.
3. Managers must generate all of the conditions and (I) They require less programming time for large simulations.
constraints for solutions that they want to examine. (2) They are usually more efficient and easier to check for
4. Each solution model is unique. Its solutions and inferences errors.
are usually not transferable to other problems. (3) They have random number generators already built-in as
subroutines.
4. Simulated average demand is based on the simulation model
constructed. is affected by the number of tIials or repetitions. and is 11. The results would change. and perhaps significantly. if a
affected by the random numbers selected. For small nu mbers of longer time period was simulated. The IO-day simulation provides
repetitions. simulated average demand can be quite variable. only an illustration of the appropriate technique. not a valid
Expected average demand is based on the mathematical model of forecast of the results of the ordering policy.
demand. and is a precise and unchanging value. the weighted aver­ 12. A computer is necessary for three reasons:
age or expectation of the values of the variable being modeled. • It can perform the individual trials in much less time than
Simulated average demand approaches expected average demand would be possible by hand.
as the number of simulation repetitions increases. • It can be programmed to examine and allow change in the
5. The role of random numbers in simulation is to help generate complex relationships being studied.
outcomes for random variables. Each random number represents a • It can generate the required sequences of random numbers.
particular possibility. 13. Illventorv orderillg policy:
6. Results of a simulation differ from run to run because differ­ • May require simulation if lead time and daily demand are
ent random numbers are selected, yielding diffe rent outcomes. If not constant. Also useful if data do not follow a traditional
the runs are very long (which is recommended). the solutions probability distribution.
should be very similar.

342
QUANTITATIVE MODULE F SIMULATION 343

Ship docking ill a pori 10 unload: Number RN

F.S
• If arrivals and unloadings do not follow Poisson/exponential of Kits Frequency Probability Interval

distributions common to queuing problems, or if other


4 4 0.10 01-10
queuing model assumptions are violated (for example, if
5 6 0.15 11-25
FIFO is not observed). 6 10 0.25 26-50
Bank leller service windows: 7 12 0.30 51-80
8 8 0.20 81-00
• If arrivals or service times do not follow standard distribu­
tions, or if several waiting lines exist, it may be easier to RN: I I 52 59 22 03 03 50 86 85 15 32 47
use simulations. Kits Usage: 5 7 7 5 4 4 6 8 8 5 6 6
u.s. economy: The average usage is: 71112 = 5.92 units
• Because mathematical equations and relationships may be
too complex to solve in closed form , and because there
F.6 At 1=
At 1 = 1
° °
RN = 07 so arrivals
RN = 60 so 1 arrival
may be random inputs or responses within the economy. Server RN = 77 and it takes 3 minutes to serve
At 1= 2 RN = 49 so I arrival
Server RN = 76 so it takes 3 minutes to serve
END-OF-MoDULE PROBLEMS
At 1=3 RN = 95 so 2 arrivals
F.l Tuna Sales RN Interval Server RN = 51 so it takes 3 minutes to serve one
01-40

Server RN = 16 so it takes 2 minutes to serve the


8
9 41-70
other one
10 71-90
Note: All checkouts are busy, so one customer waits.
11 91-00

°
At t = 4 RN = 14 so arrivals
Therefore, at the end of 5 minutes two checkouts are
Sales: 8,9, 10,9,9, 9, 8, 8, 10, 10
still busy and one is available.
F.2 Breakdowns RN Interval
F.7 Random number
o 01-50

1 51-80
intervals
2 81-00
01-01
02-16
Breakdowns: 0,0, 0,0, 0, 0,0, 2, 0,2: Proportion = 20% 17-46
47-61
F.3 Here is a table showing the service now: 62-65
66-75
Customer Arrival Service Service Service Time in Time in 76-00
Number Time Time Begins Ends Line System
91 generates demand of 220; 45 and 37 each generate
8:01 6 8:01 8:07 0 6
demand of 140; 65 generates 180; and 51 generates 160.
2 8:06 7 8:07 8:14 1 8
3 8:09 8 8:14 8:22 5 13
Their sum is 840, and their average is 168.
4 8:15 6 8:22 8:28 7 13
5 8:20 6 8:28 8:34 8 14

(a) 8: 14 Service starts for customer #3


(b) 8:34 Customer #5 leaves
(c) 4.25 min. average time in queue (= 21/5)
(d) 10.8 min. average time in system

F.4 Day RN Demand Unsold Lost Profit


1 52 200 0 0 20.00
2 06 100 100 0 -15.00
3 50 200 0 0 20.00
4 88 250 0 50 17.50
5 53 200 0 0 20.00
Average profit = 12.50
344 QUANTITATIVE MODULE F SIMULATION

F.8 Time Between


Arrivals Prob. RN Interval Service Time Prob. RN Interval
1 0.2 01-20 1 0.3 01-30
2 0.3 21-50 2 0.5 31-80
3 0.3 51-80 3 0.2 81-00
4 0.2 81-00

First arrival (RN = 14) at 11 :01. Service time 3 (RN


= 88)
Leaves at 11 :04.

Second arrival (RN = 74) at J 1:04 (J minutes after


1st).
Service time = 2 (RN = 32). Leaves at 11:06.

Third ani val (RN= 27) at I I :06. Service time = 2 (RN = 36).

Leaves at 11 :08.

Fourth arrival (RN = 03) at 11 :07. Must wait I minute for

service to start. Service time = I minute (RN = 24). Leaves at


II :09.

F.9 Number of failed


Boxes per month Probability RN Intervals
0 0.10 01-10
1 0.14 11-24
2 0.26 25-50
3 0.20 51-70
4 0.18 71-88
5 0.12 89-00

No. of Three Month


No. of Three Month
Month RN Failures Total
Month RN Failures Total
1 37 2 13 41 2 7
2 60 3 14 31 2 8
3 79 4 9
15 97 5 9
4 21 8
16 78 4 11
5 85 4 9
17 94 5 14
6 71 4 9
18 44 2 11
7 48 2 10
19 32 2 9
8 39 2 8
20 85 4 8
9 31 2 6
21 64 3 9
10 35 2 6
22 84 4 11
11 12 1 5
23 63 3 10
12 73 4 7
24 29 2 9

The average number of f ailures is 69/24 = 2.88 units per

month.

There is a period, over months 9-11 , during which less than

7 units were demanded over each three-month stretch.

F.I0 (a) Number of Freq. Probability Cumulative Random No. Interval


(b)
3 or less 0 0.00 0.00
4 20 0.10 0.10 01-10
5 30 0.15 0.25 11-25
6 50 0.25 0.50 26-50
7 60 0.30 0.80 51-80
8 40 0.20 1.00 81-00
9 0
.E = 200 .E= 1.00
QUANTITATIVE MODULE F SIMULATION 345

(c) Hour Random* Arrivals


52 7

2 37 6

3 82 8

4 69 7

5 98 8

6 96 8

7 33 6

8 50 6

9 88 8

10 90 8

11 50 6

12 27 6

13 45 6

14 81 8

15 66 7

L' = 105

* Random numbers taken from Col umn I of Table F.4, starting


at the top.
. 105
Average number of am vals per hour = - = 7 cars
15
F .Il (a) Day 3 demand = 24
(b) Net profit total = $36.70
(c) Lost goodwill on day 6 = $.30
(d) Net profit on day 2 = $6.10
(e) Papers ordered day 5 = 21

Demand p Cum. prob. R.N.lnt.


21 0.25 0.25 01-25
22 0.15 0.40 26-40
23 0 .10 0.50 41-50
24 0.20 0.70 51-70
25 0.30 1.00 71-100

Papers Random Goodwill


Day ordered number Demand Revenue Cost cost Net profit
1 22 37 22 $11.00 $4.40 $0.00 $6.60
2 22 19 21 $10.50 $4.40 $0.00 $6.10
3 21 52 24 $10.50 $4.20 $0.30 $6.00
4 24 8 21 $10.50 $4.80 $0.00 $5.70
5 21 22 21 $10.50 $4.20 $0.00 $6.30
6 21 61 24 $10.50 $4.20 $0.30 $6.00

F.12 Demand for Cumulative Random Number


Ace Drill Freq. Probability Probability Interval
0 15 0.05 0.05 01-05
30 0.10 0.15 06-15
2 60 0 .20 0.35 16-35
3 120 0.40 0.75 36-75
4 45 0.15 0.90 76-90
5 30 0.10 1.00 91-00
L'= 300
346 QUANTITATIVE MODULE F SIMULATION

lead Time Cumulative Random Number

(Da~s) Freg. Probabilit~ Probabilit~ Interval

10 0.20 0.20 01-20

2 25 0.50 0.70 21-70

3 15 0.30 1.00 71-00

I=50

Time Beg Rand End lost Place Rand lead

Period Inv Num* Demand Sold Inv Sale Order Num Time

1 12 07 1 1 11 0 No

2 11 60 3 3 8 0 No

3 8 77 4 4 4 0 Yes 49 2

4 4 76 4 4 0 0 No

5 0 95 5 0 0 5 No

6 12 51 3 3 9 0 No

7 9 16 2 2 7 0 No

8 7 14 1 1 6 0 Yes 85 3

9 6 59 3 3 3 0 No

10 3 85 4 3 0 1
-

No

I=30 24 48 6

* Random numbers taken from the right-hand colullln of Table F.4, F.14 Heater Probability Random No. Interval
reading top-to-bottolll, in order used . 3 0.02 01-02
Daily order cost = $10.00 x 2/ I 0 = $2.00 4 0.09 03-11
5 0.10 12-21
Daily holding cost = $0.50 x 48/1 0 = $2.40
6 0.15 22-36
Daily stockout cost = $8.00 x 6/ I 0 = $4.80
7 0.25 37-61
$9.20 62-73
8 0.12
The cost is greater ($9.20 vs. $6.(5) than that found with 9 0.12 74-85
the order quantity of 10 units. Note, however, that the short 10 0.10 86-95
time period simulated does not permit a valid analysis of 11 0.05 96-00
either alternative. 1.00

F.13 Selling price = $2 (a) Week Random Simulated Sales


Cost = $0.80
1 10 4
(a) Cumulative Random 2 24 6
Demand Probability probability number 3 03 4
2,300 0.15 4 32 6
0.15 1-15
5 23 6
2,400 0.22 0.37 16-37
6 59 7
2,500 0.24 0.61 38-61
7 95 10
2,600 0.21 0.82 62-82
8 34 6
2,700 0.18 1.00 83-00
9 34 6
10 51 7
Random Produce 2,500 Produce 2,600 11 08 4
Number Demand Profit Profit 12 48 7
7 2,300 $2,600 $2,520 13 66 8
60 2,500 3,000 2,920 14 97 11
77 2,600 3,000 3,120 15 03 4
49 2,500 3,000 2,920 16 96 11
76 2,600 3,000 3,120 17 46 7
95 2,700 3,000 3,120 18 74 9
51 2,500 3,000 2,920 19 77 9
16 2,400 2,800 2,720 20 44 7
14 2,300 2,600 2,520
139
85 2,700 3,000 3,120

With a supply of 8 healers, Higgins will stock


(b) [I' the company produces 2.500 programs, the average out 5 times during the 20-week period (in weeks
profit is $2,900. 7,14,16, 18, and 19).
(c) If the company produces 2.600 programs, the average

profit is also $2,900.

QUANTITATIVE MODULE F SIMULATION 347

(b) Average sales as given by the results of the simulation:


Total sales 139
A verage sales = =­ = 6.95 per week
20 20
Other simulations by students may yield slightly
different results.
(c) Using expected values:
£(sales) = 0.02 x 3 + 0.09 x 4 + 0.10 x 5
+ 0.15 x 6 + 0.25 x 7 + 0.12 x 8
+ 0.12 x 9 + 0.10 x 10 + 0.05 x 1 I
= 7.16 heaters

A si mulation that lasts for longer than 20 time periods


will result in answers that are even closer.

F.lS The Monte Carlo simulation tables are:

Time between
Arrivals RN Time in Tanning RN
(minutes) Probability Interval Bed (minutes) Probability Interval
5 0.30 01-30 10 0.20 01-20
10 0.25 31-55 15 0.30 21-50
15 0.20 56-75 20 0.40 51-90
20 0.15 76-90 25 0.10 91-00
25 0.10 91-00

The following table shows the use of Table F.4. The 14th
column was used for arrivals and the 8th column was used
for tanning times.

Arrivals Training Time Wait Time


Time
Customer RN Arrived RN Time Begin End Begin End Balk?
0 2:00 20 2:00 2:20
1 32 2:10 47 15 2:20 2:35 2:10 2:20
2 73 2:25 03 10 2:35 2:45 2:25 2:35
3 41 2:35 11 10 2:45 2:55 2:35 2:45
4 38 2:45 10 10 2:55 3:05 2:45 2:55
5 73 3:00 67 20 3:05 3:25 3:00 3:05
6 01 3:05 23 15 3:25 3:40 3:05 3:25
7 09 3:10 89 20 Yes
8 64 3:25 62 20 3:40 4:00 3:25 3:40
9 34 3:35 56 20 Yes
10 55 3:45 74 20 4:00 4:20 3:45 4:00
11 84 4:05 54 20 4:20 4:40 4:05 4:20
12 16 4:10 31 15 Yes
13 98 4:35 62 20 4:40 5:00 4:35 4:40
14 49 4:45 37 15 5:00 5:15 4:45 5:00
15 00 5:10 33 15 5:15 5:30 5:10 5:15
16 30 5:15 33 15 5:30 5:45 5:15 5:30
17 23 5:20 82 20 Yes
18 46 5:30 68 20 5:45 6:05 5:30 5:45
19 75 5:45 50 15 6:05 6:20 5:45 6:05
20 13 5:50 22 15 Yes
21 29 5:55 27 15 Yes ·
22 08 6:00 08 10 Yes

During these four hours, 7 customers balked at waiting for a


bed. The total time that Taboo missed in terms of payments
was 105 minutes. This equates to 105/10 x $2.00 or $21.00
per day. This translates to $504 per month. The addition of
another tanning bed is not justified.
348 QUANTITATIVE MODULE F SIMULATION

F.16 Demand for Cumulative Random Number


Mercedes Freq. Probability Probability Interval*
6 3 0.083 0.083 01-08
7 4 0.111 0.194 09-19
8 6 0.167 0.361 20-36
9 12 0.333 0.694 37-69
10 9 0.250 0.944 70-94
11 0.028 0.972 95-97
12 0.028 1.000 98-00
L=36 L= 1.000

* Note that the cumulative probabilities have been rounded to two


significant digits when used to develop the random number
intervals.

Lead Probability Cumulative Random No. Interval


1 0.44 0.44 01--44
2 0.33 0.77 45-77
3 0.16 0.93 78-93
4 0.07 1.00 94-00
We have arbitrarily chosen a beginning inventory of 14 cars.

Time Beg Rand Demand Sold End Lost Place Rand Lead
1 14 07 6 6 8 0 Yes 60 2
2 8 77 10 8 0 2 No
3 0 49 9 0 14 9 No
4 14 76 10 10 4 0 Yes 95 4
5 4 51 9 4 0 5 No
6 0 16 7 0 0 7 No
7 0 14 7 0 0 7 No
8 0 85 10 0 14 10 No
9 14 59 9 9 5 0 Yes 85 3
10 5 40 9 5 0 4 No
11 0 42 9 0 0 9 No
12 0 52 9 0 14 9 No
13 14 39 9 9 5 0 Yes 73 2
14 5 89 10 5 0 5 No
15 0 88 10 0 14 10 No
16 14 24 8 8 6 0 Yes 01
17 6 11 7 6 14 1 No
18 14 67 9 9 5 0 Yes 62 2
19 5 51 9 5 0 4 No
20 0 33 8 0 14 8 No
21 14 08 6 6 8 0 Yes 40
22 8 29 8 8 14 0 No
23 14 75 10 10 4 0 Yes 33
24 4 95 -
11 4 -
14 7 No
L 209 112 157 97 16

Useful statistics from the simulation:


• A verage demand
Simulation-t209/24 = 8.71
Theoretical 8.75
• A verage lead ti me
Simulation-t 16/8 = 2.00
Theoretical 1.86
• Average ending inventory = 157/24 = 6.5
• Average number of lost sales = 97124 = 4.04
QUANTITATIVE MODULE F SIMULATION 349

F.17 C, = carry cost 24 x 600 x 6.50 = 93600


+ lost sale cost 4350 x 97 = 421950
+ order cost 8 x 570 4560
520110
= $520,110, or $21 ,671 per month
F.18 As in problem F.16, we have arbitrarily chosen a beginning
inventory of 14 cars. (See simulation below.)
Useful statistics from the simulation:
• A verage demand
Simulation~20l/24 = 8.38
Theoretical 8.75
• Average lead time
Simulation~ 13/6 = 2.17
Theoretical 1.86
• Average ending inventory = 214/24 = 8.92
• Average number of lost sales = 82/24 = 3.42
C, = catTy cost 24 x 600 x 8.92 = 128448
+ lost sale cost 4350 x 82 = 356700
+ order cost 6 x 570 3420
488568
It would appear the inventory policy of ordering in quanti­
ties of 21 using a reorder point of 10 is to be preferred.
Order quantity = 14, reorder point = 14
~$520, II 0 or $21 ,671 Imonth
Order quantity = 21, reorder point = 10
~$488,568 or $20,357/month

Students should be cautioned that the single 24-period


simulation of each condition does not provide sufficient
information upon which to base any firm conclusion.
It would also be useful, at this point, to discuss with
students some of the considerations with regard to choosing
random numbers in a way as to promote comparability of
the two simulations.

Time Beg Rand Demand Sold End Lost Place Rand Lead

14 07 6 6 8 0 Yes 60 2
2 8 77 10 8 0 2 No
3 0 49 9 0 21 9 No
4 21 76 10 10 11 0 No
5 11 59 9 9 2 0 Yes 51 2
6 2 16 7 2 0 5 No
7 0 14 7 0 21 7 No
8 21 85 10 10 11 0 No
9 11 59 9 9 2 0 Yes 85 3
10 2 40 9 2 0 7 No
11 0 42 9 0 0 9 No
12 0 52 9 0 21 9 No
13 21 39 9 9 12 0 No
14 12 73 10 10 2 0 Yes 89 3
15 2 88 10 2 0 8 No
16 0 24 8 0 0 8 No
17 0 01 6 0 21 6 No
18 21 11 7 7 14 0 No
19 14 67 9 9 5 0 Yes 62 2
20 5 51 9 5 0 4 No
21 0 03 6 0 21 6 No
22 21 08 6 6 15 0 No
23 15 40 9 9 6 0 Yes 29
24 6 29 8 6 21
-2 No
Sums 201 119 214 82 13
350 QUANTITATIVE MODULE F SIMULATION

F.19 (a) The Appropriate random number table are: Customer Service Times

Interarrival Times for 500 observation Ala Carte Value


Service Meals
Times between Arrivals Number of Random
Times Random Service Random
(in minutes) Occurrences Numbers
(mins.) Numbers Times (mins.) Numbers
1 100 01-20 01-20
01-10
2 150 21-50
2 11-35 2 21-55
3 125 51-75 36-70 56-80
3 3
4 100 76-95
4 71-00 4 81-00
5 25 96-00
We must use three random number streams. The first is
Type of Meals Ordered used to determine the time of the next arrival. The
second is used to determine if that arrival will order Ala
Meal Type Probability Random Number
Carte or a value meal. The last stream is used to
Ala Carte 0.75 01-75 determine the service time for the order.
Value Meals 0.25 76-00 The last three columns of Table F.4 have been
chosen. Different stream will result in different out­
comes. The instructor may want to average the results of
the class to obtain one large simulation result.

Arrivals Meal Selection Service Times


Rand Arrival No. in Rand Meal Rand Service Begin End
No. Time Queue Balk? No. Type No. Time Time Time
84 11 :04 0 No 57 Ala 07 11 :04 11 :05
55 11 :07 0 No 17 Ala 60 3 11 :07 11: 10
25 11 :09 No 36 Ala 77 4 11: 10 11: 14
71 11: 12 No 72 Ala 49 3 11: 14 11: 17
34 11: 14 1 No 85 Value 76 3 11: 17 11 :20
57 11: 17 1 No 31 Ala 95 4 11 :20 11:24
50 11: 19 2 No 44 Ala 51 3 11 :24 11 :27
44 11 :21 2 No 30 Ala 16 2 11 :27 11 :29
95 11 :25 2 No 26 Ala 14 2 11 :29 11 :31
64 11 :28 2 No 09 Ala 85 4 11 :31 11 :35
16 11 :29 2 No 49 Ala 59 3 11 :35 11 :38
46 11 :31 2 No 13 Ala 85 4 11 :38 11 :42
54 11 :34 3 No 33 Ala 40 3 11 :42 11 :45
64 11 :37 3 No 89 Value 42 2 11 :45 11 :47
61 11 :40 3 No 13 Ala 52 3 11 :47 11 :50
23 11 :42 3 No 37 Ala 39 3 11 :50 11 :53
01 11 :43 4 Yes 58 73
79 11 :47 2 No 45 Ala 89 4 11:53 11 :57
19 11 :48 3 No 95 Value 88 4 11 :57 12:01
50 11 :50 3 No 33 Ala 24 2 12:01 12:03
05 11 :51 4 Yes 75 01
86 11 :55 3 No 29 Ala 11 2 12:03 12:05
62 11 :58 3 No 40 Ala 67 3 12:05 12:08
22 12:00 4 Yes 08 62

The table shows that three meals were lost. They would
have been Ala Carte. This is 3 x $3 = $9 per hour and
$18 per day (a day being the IIAM-lpM lunch hour
period). For 20 days. the lost revenue is $360.

Number of Ala Carte Meals Served =18


Value of Ala Carte Meals Served = 18 x $3 = $54
Number of Value Meals Served = 3
Value of Value Meals Served = :1 x $2.25 = $6.75
Total Contribution for One Hour = $60.75
.... :;:':,~;::;.-

QUANTITATIVE MODULE F SIMULATION 351

(b) If the price of the meals change, the only random number
table that is altered is the Type of Meals Selected. The new
table is:

Type of Meals Ordered


Meal Type Probability Random Number

Ala Carte 0.60 01-60


Value Meals 0.40 61-00

The new outcome are shown in the following table:

Arrivals Meal Selection Service Times

Rand Arrival No. in Rand Meal Rand Service Begin End


No. Time Queue Balk? No. Type No. Time Time Time
84 11 :04 0 No 57 Ala 07 1 11 :04 11 :05
55 11 :07 0 No 17 Ala 60 3 11 :07 11: 10
25 11 :09 No 36 Ala 77 4 11: 10 11 :14
71 11 :12 No 72 Value 49 2 11 :14 11: 16
34 11 :14 No 85 Value 76 3 11: 16 11: 19
57 11 :17 No 31 Ala 95 4 11: 19 11 :23
50 11: 19 1 No 44 Ala 51 3 11 :23 11 :26
44 11 :21 2 No 30 Ala 16 2 11 :26 11 :28
95 11 :25 2 No 26 Ala 14 2 11:28 11 :30
64 11 :28 No 09 Ala 85 4 11 :30 11 :34
16 11 :29 2 No 49 Ala 59 3 11 :34 11 :37
46 11 :31 2 No 13 Ala 85 4 11 :37 11 Al
54 11 :34 2 No 33 Ala 40 3 11 Al 11 A4
64 11 :37 2 No 89 Value 42 2 11 A4 11 A6
61 11 AO 3 No 13 Ala 52 3 11 A6 11 A9
23 11 A2 3 No 37 Ala 39 3 11 :49 11 :52
01 11 A3 4 Yes 58 73
79 11 :47 2 No 45 Ala 89 4 11 :52 11 :56
19 11 A8 3 No 95 Value 88 4 11 :56 12:00
50 11 :50 3 No 33 Ala 24 2 12:00 12:02
05 11 :51 4 Yes 75 01
86 11 :55 3 No 29 Ala 11 2 12:02 12:04
62 11 :58 3 No 40 Ala 67 3 12:04 12:07
22 12:00 3 No 08 Ala 62 3 12:07 12:10

The number of missed sales dropped from 3 to 2. This is a


saving of $3.00
Number of ala Carte Meals Served = 18
Value of ala Carte Meals Served = 18 x $3 = $54
Number of Value Meals Served = 4
Value of Value Meals Served = 4 x $2.00 = $8.00
Total Contribution for One Hour = $62.00
While we have gained one Value meal at a profit of $2.00,
this amount is offset by the loss of 75 cents on the three value
meals we were already providing. The net gain over one
month will be $1.25 x 2 x 20 = $45. The gain may not be
worth the pain association with making new signs and
changing the pIices in the cash register.
352 QUANTITATIVE MODULE F SIMULATION

F.20 Random
Cumulative Number
From To Probability Probability Interval
Initial Exam X-ray Dept. 0.45 0.45 01-45
Operating Room 0.15 0.60 46-60
Observation 0.10 0.70 61-70
Out-Processing 0.30 1.00 71-00
X-Ray Dept. Operating Room 0.10 0.10 01-10
Cast-Fitting 0.25 0.35 11-35
Observation 0.35 0.70 36-70
Out-Processi ng 0.30 1.00 71-00
Operating Room Cast-Fitting 0.25 0.25 01-25
Observation 0.70 0.95 26-95
Out-Processing 0.05 1.00 96-00
Cast-Fitting Observation 0.55 0.55 01-55
X-ray Dept. 0.05 0.60 56-60
Out-Processing 0.40 1.00 61-00
Observation Operating Room 0.15 0.15 01-15
X-ray Dept. 0.15 0.30 16-30
Out -Processi ng 0.70 1.00 31-00

(a) Simulation

Random
Person Number* From To
Person 1 52 Initial Exam Operating Room
37 Operating Room Observation
82 Observation Out-Processing
Person 2 69 Initial Exam Observation
98 Observation Out -Processi ng
Person 3 96 Initial Exam Out -Processi ng
Person 4 33 Initial Exam X-Ray Dept.
60 X-Ray Dept. Observation
88 Observation Out-Processing
Person 5 90 Initial Exam Out-Processi ng
Person 6 50 Initial Exam Operating Room
27 Operating Room Observation
45 Observation Out-Processing
Person 7 81 Initial Exam Out-Processing
Person 8 66 Initial Exam Observation
74 Observation Out-Processing
Person 9 30 Initial Exam X-Ray Dept.
59 X-ray Dept. Observation
67 Observation Out-Processing
Person 10 60 Initial Exam Operating Room
60 Operating Room Observation
80 Observation Out-Processing

* Random number taken from Column I of lable F.4. Reading


(op-Io-bollom.

Using this small simulation, we cannot really draw any


firm conclusions. However, no one was observed to
enter the X-ray Department twice. The original proba­
bility distributions would, however. suggest that this
certainly could happen in a longer simulation.
QUANTITATIVE MODULE F SIMULATION 353

F.21 Time Between Random Number


Arrivals Probability Interval

0.20 01-20

2 0.25 21-45

3 0.30 46-75

4 0 .15 76-90

5 0.10 91-00

Random Number

Service Time Probability Interval

0.10 01-10

2 0.15 11-25

3 0.35 26-60

4 0.15 61-75

5 0.15 76-90

6 0.10 91-00

(a) Simulation of a one-teller drive-through:

Arrivals Service
Rand Time Till Arrive at Rand Time Wait Time
Num Next Time Num Needed Begin End (Mins)
52 3 1 :03 60 3 1 :03 1:06 0

37 2 1 :05 60 3 1 :06 1:09 1

82 4 1 :09 80 5 1 :09 1 :14 0

69 3 1: 12 53 3 1 :14 1: 17 2

98 5 1: 17 69 4 1 :17 1:21 0

96 5 1 :22 37 3 1 :22 1 :25 0

33 2 1:24 06 1 :25 1 :26 1

50 3 1:27 63 4 1:27 1:31 0

88 4 1:31 57 3 1:31 1:34 0

90 4 1 :35 02 1 1 :35 1 :36 0

50 3 1 :38 94 6 1 :38 1 :44 0

27 2 1 :40 52 3 1 :44 1 :47 4

45 2 1 :42 69 4 1 :47 1:51 5

81 4 1 :46 33 3 1:51 1:54 5

66 3 1 :49 32 3 1 :54 1 :57 5

74 3 1 :52 30 3 1 :57 2:00 5

30 2 1 :54 48 3 2:00 2:03 6

59 3 1 :57 88 5 2:03 2:08 6

67 3 2:00 .[=40

Yearly waiting costs are given by:

Waiting cost = (40 min/hr)(7 hr/day)(200 days)($l Imin)

= $56,000
~-::,.;,>

354 QUANTITATIVE MODULE F SIMULATION

(b) Simulation of two one-teller drive-throughs

Arrivals Service
Time Teller 1 Teller 2 Wait
Rand Till In At Rand Time Time
Num Next Time Num Req Beg End Beg End (Mins)

52 3 1:03 60 3 1:03 1:06 0


37 2 1:05 60 3 1:05 1:08 0
82 4 1:09 80 5 1:09 1:14 0
69 3 1: 12 53 3 1:12 1:15 0
98 5 1: 17 69 4 1:17 1:21 0
96 5 1:22 37 3 1:22 1:25 0
33 2 1:24 06 1 1:24 1:25 0
50 3 1:27 63 4 1:27 1:31 0
88 4 1:31 57 3 1:31 1:34 0
90 4 1:35 02 1 1 :35 1:36 0
50 3 1:38 94 6 1:38 1:44 0
27 2 1:40 52 3 1:40 1:43 0
45 2 1:42 69 4 1:43 1:47 1
81 4 1:46 33 3 1:46 1:49 0
66 3 1:49 32 3 1:49 1:52 0
74 3 1:52 30 3 1:52 1:55 0
30 2 1:54 48 3 1:54 1:57 0
59 3 1:57 88 5 1:57 2:02 0
67 3 2:00 2=1

Wailing cost = (1 min/hr) (7 hrs/day) (200 days) ($I/min)


= $1 ,400
(c) Cost alternatives:
Wait Drive-through Labor (Teller)
Cost/year = + +
Cost/year Amortization/year Cost/year
Cost for single one-teller drive-through:
$56,000 + $12,000 + $16,000 = $84,000
Cost for two one-teller drive-throughs:
$1,400 + $20,000 + $32,000 = $53,400
Cost savings achieved by using two booths:
$84,000 -$53,400 = $30,600
The conclusion is to place two teller booths in use. It is
criticaL however, that the student realize that the simulation
given above should be replicated for a much longer time
period before one draws any firm conclusions.

F.22 (a) Profit = (amount produced)(sales price)


- (ingredient I cost)(ingredient I units)
- (ingredient 2 cost)(ingredient 2 units)
= 30(sales price) - $50(25 units)
- (ingredient 2 cost)(36 units)
= 30(sales price) - $1,250
- 36(ingredient 2 cost)
where sales price and ingredient 2 cost are probabilistic.
(b) Expected sales price = 0.2 ($300) + 0.5($350) + 0.3($400)
= $355
Expected ingredient 2 cost = 0.1 ($35) + 0.6($40) + 0.3($45)
=$41
Expected profit = 30($355) - $1 ,250 - 36($41)
= $7,924/day
QUANTITATIVE MODULE F SIMULATION 355

(C) Daily
Random Sales Gross Random Ingred.2 Ingred.2 Ingred.
Day Number Price Sales Number Cost/Unit Cost Total 1 Cost Profit
1 52 $350 10,500 37 $40 $1,440 $1,250 $7,810
2 06 300 9,000 66 40 1,440 1,250 6,310
3 50 350 10,500 91 45 1,620 1.250 7,630
4 88 400 12,000 35 40 1,440 1,250 9,310
5 53 350 10,500 32 40 1,440 1,250 7,810
6 30 350 10,500 00 45 1,620 1,250 7,630
7 10 300 9,000 84 45 1,620 1,250 6,130
8 47 350 10,500 57 40 1,440 1,250 7,810
9 99 400 12,000 07 35 1,260 1,250 9,490

Random number intervals for sales price: Random number intervals for cost 2:
01 - 20 = $300 01 - 10=$35
21-70 = $350 11 - 70 = $40
71-00= $400 71 - 00=$45
(d) Expected profit from simulation = $7,770/day

INTERNET HOMEWORK PROBLEMS F.2S Sales RN Interval Lead Time RN Interval


---------------­
Solutions to problems on our companion web site 7 01-30 3 01-20
(www.prenhall.comlheizel·). 8 31-70 4 21-80
9 71-00 5 81-00
F.23 Demand RN Interval
First order: RN = 60 so lead time = 4 days.
4 01-10
5 11--40 Demand day I 8 (RN = 52)
6 41-70 day 2 9 (RN = 78)
7 71-90 day 3 7 (RN= 13)
8 91-00 day 4 7 (RN = 06)
Random number stream: 88 02 28 49 36 87 21 95 50 24 Total demand during lead time = 31. Because the reorder
Demand: 7 4 5 6 5 7 5 8 6 5 point is 32, there is no stockout. Second order: RN = 87
F'.24 so lead time = 5
nCars RN Interval
6 01-20 Demand day 1 9 (RN = 99)
7 21-44 day 2 9 (RN = 98)
8 45-84 day 3 9 (RN = 80)
9 85-00 day 4 7 (RN = 09)
day 5 8 (RN = 67)
Arrivals: 7,7,7,6,8,6, 7, 9,6,7; Average = 70/JQ = 7 Total demand during lead time = 42. So the company
experienced a stockout during this time.
356 QUANTITATIVE MODULE F SIMULATION

F.26 Number of Air Relative Frequency Cumulative Random Number


Conditioner Failures (Probability) Probability Interval
0 0.06 0.06 01-06
1 0.13 0.19 07-19
2 0.25 0.44 20--44
3 0.28 0.72 45-72
4 0.20 0.92 73-92
5 0.07 0.99 93- 99
6 0.01 1.00 00

Number of AIC Compressors


Simulated Period Random Number* to Fail This Year
1 50 3
2 28 2
3 68 3
4 36 2
5 90 4
6 62 3
7 27 2
8 50 3
9 18 1
10 36 2
11 61 3
12 21 2
13 46 3
14 01 0
15 14
16 81 4
17 87 4
18 72 3
19 80 4
20 46 3

* Random numbers taken from Column :I of Table FA. stat1ing at


the top.
No, it is not common to find three or more years in a row with
two or less compressor failures.

F.27 Cumulative Random Number


Number of Arrivals Probability Probability Interval
0 0.13 0.13 01-13
1 0.17 0.30 14-30
2 0.15 0.45 31--45
3 0.25 0.70 46-70
4 0.20 0.90 71-90
5 0.10 1.00 91-00

Cumulative Random Number


Daily Uploading Rate Probability Probability Interval
1 0.03 0.03 01-03
2 0.12 0.15 04-15
3 0.40 0.55 16-55
4 0.28 0.83 56-83
5 0.12 0.95 84-95
6 0.05 1.00 96-00
QUANTITATIVE MODULE F SIMULATION 357

Number Daily Total to be Number


Day Delayed Random* Arrivals Unloaded Random** Unloaded
1 37 2 2 69 2
2 0 77 4 4 84 4
3 0 13 0 0 12 0
4 0 10 0 0 94 0
5 0 02 0 0 51 0
6 0 18 1 1 36 1
7 0 31 2 2 17 2
8 0 19 1 1 02
9 0 32 2 2 15 2
10 0 85 4 4 29 3
11 31 2 3 16 3
12 0 94 5 5 52 3
13 2 81 4 6 56 4
14 2 43 2 4 43 3
15 31 2 3 26 3
I =6 I=31 I= 31
* Random numbers taken from the hottom row of Table F.4, Month Beg Rand* Income Rand** Expense End
reading left-to-right.
** Random numbers taken from the next-to-Iast row of Table F.4, 600 52 400 37 400 600
reading left-to-right. 2 600 06 350 63 500 450
3 450 50 400 28 400 450
6 4 600 88 450 02 300 600
Average number of barges delayed per day = - = 0,4
15 5 500 53 400 74 500 500
6 450 30 350 35 400 450
,
Average amva1s per day = -31 = 20
, 7 7 400 10 350 24 400 400
15
8 400 47 400 03 300 500
31 9 500 99 500 29 400 600
Average unloaded per day = - = 2.07
15 10 600 37 350 60 500 450
11 450 66 450 74 500 400
(b) The short time span simulated (15 days) introduces
12 400 91 500 85 500 400
volatility in both the daily arrival rate (from 2.73
arrivals per day in the original example to only 2.07 in * Random numbers taken from the top row of Table F.4, reading
this modified simulation). This, coupled with speedier left-to-right.
unloading rate, produces a much lower average delay ** Random numbers taken from the second row of Table F.4,
rate (from 1.33 per day down to only 0,4 per day). reading left-to-right.

F.28 Monthly Probability Cumulative Random No. Interval Her balance never drops below $400, and she should be
able to balance her account.
350 0.40 0.40 01-40
400 0.20 0.60 41-60
450 0.30 0.90 61-90
500 0.10 1.00 91-00

Monthly Probability Cumulative Random No. Interval


300 0.10 0.10 01-10
400 0.45 0.55 11-55
500 0.30 0.85 56-85
600 0.15 1.00 86-00
358 QUANTITATIVE MODULE F SIMULATION

F.29 MTBF = Mean Time Between Failures For four pens replaced:
(a) Simulation Approach £(MTBF) = 100 x 0.15 + 110 x 0.25 + 120 x 0.35
MTBF Random + 130 x 0.20 + 140 x 0.05 = I 17.5 hours
One Pen Cumulative Number Cost/breakdown = 4 x $8/ pen + $1 OO/repai r = $132
Replaced Probability Probability Interval C" = $132/breakdown -;- I 17.5 hours/breakdown
10 0.05 0.05 01-05 = $1.12/hour
20 0.15 0.20 06-20
30 0.15 0.35 21 - 35 Cost
40 0.20 0.55 36-55
Method Replace One Pen Replace Four Pens
50 0.20 0.75 56-75
60 0.15 0.90 76-90 Simulation $1 .53/hr $1.09/hr
70 0.10 1.00 91-00 Analytical $1.38/hr $1.12/hr

In either case. the preferred approach appears to be to


MTBF All Random replace the four pens.
Four Pens Cumulative Number
Replaced Probability Probability Interval
CASE STUDY
100 0.15 0.15 01-15
110 0.25 OAO 16-40 ALABAMA AIRLINES' CALL CENTER
120 0.35 0.75 41-75
This case describes a single- server queuing scenari o that is to bc
130 0.20 0.95 76-95
investigated by simulation. The basic premise can be applied to
140 0.05 1.00 96-00
various queuing scenarios (e.g .. bank. post office. cinema box of­
fice. among others) and i s not restricted to the Alabama Airlines '
One Pen Replaced Four Pens Replaced reservation scenario. The purpose of the case is to introduce
Random Random students to simulation as a powerful decision-making aid.
Number* MTBF (Hours) Number** MTBF (Hours) When presenting simulation modeling techniques in the
47 40 99 140 classroom , one is faced with the following dilemma: the trade-off
03 10 29 110 between the ben efits of the modeling and decision-making process
11 20 27 110 versus the complexity and the tedium of the required calculations.
10 20 75 120 Spreadsheets that have a random number generator, such as Excel ,
67 50 89 130 provide a means whereby th e principles and methodology of
23 30 78 130 simulation can be introduced and illustrated by investigating th e
89 60 68 120 decision-making process for meaningful size problems without
62 50 64 120
making the associated calculation requirements too burdensome.
56 50 62 120
A spreadsheet. su ch as ExeeL is envisaged as the vehicle for this
74 50 30 110
simulation.
I= 380 I = 1210
In the assignment the students consider themsel ves to be op­
* Random numbers taken from Column 8 of Table F.4, readin g erations management consultants investigating the particular sce­
top-to-bottom. nario for their client(s). The assignments are presented in the form
** Random numbers taken from Column 9 or Table F.4, of a report to their clients. For each of the scenarios, an "executive
reading top-to-bottom.

When one pen is replaced :

c, = 10 pens x $8 + 10 repairs x $50 / hr x I hr/repair = $580

C, 580 $ .
C, = = -= 1.53/hl

, 380 hours 380

When four pens are replaced:

C, = 40 pens x $8 + 10 repairs x $50/ hr x 2 hrs/repair = $1320


C 1320
C = I = - - = $1.09/hr

b 1210 hours 1210

(b) AnalYlical Approach


For one pen replaced:

£ (MTBF) = 10 x 0.05 + 20 x 0.15 + 30 x 0.15 + 40 x 0.20 + 50 x 0.20 + 60 x 0.15 + 70 x O. 10


= 42 hours

Cost/breakdown = $8/pen + $50/repair = $58


C,,= $58/breakdown -;- 42 hours/breakdown = $1.38/hour
QUANTITATIVE MODULE F SIMULATION 359

summary" is required stating the main results and recommenda­


tions. The body of the report then supports these recommendations
with precise explanations, relevant data, and models. Students are
encouraged to be concise and to the point.
A spreadsheet model for the scenario should be constructed
consisting of the following items:

Arrival Interval Distribution


Random No, Range Arrival Gap
Prob. Lower Limit Upper Limit (mins)
0.11 0 10
0.21 11 31 2
0.22 32 53 3
0.20 54 73 4
0.16 74 89 5
0.10 90 99 6

Service Time Distribution


Random No. Range Service Time
Prob. Lower Limit Upper Limit (mins)
0.20 0 19 1
0.19 20 38 2
0.18 39 56 3
0.17 57 73 4
0.13 74 86 5
0.10 87 96 6
0.03 97 99 7

These above two tables are used for the random number generation
of alTival and service times to the given distributions. (Random
number lower limits could have started at 0 I, with upper limits
set at 00.)

One Trial
Time Time Time
Cust Rand Arrival Rand Service Arrival Service Service in on Server
No. No. Gap No. Time Time Start End Syst Hold Idle Utilization
Summary for this Trial Run averages: 15.5 12.0 0.15 95.9%
maximums: 34.0 32.0 2.00
0 0
27 2 82 5 2 2 7 5 0 2
2 8 60 4 3 7 11 8 4 0
3 93 6 25 2 9 11 13 4 2 0
4 93 6 36 2 15 15 17 2 0 2
5 65 4 91 6 19 19 25 6 0 2
6 44 3 88 6 22 25 31 9 3 0
7 41 3 23 2 25 31 33 8 6 0
360 QUANTITATIVE MODULE F SIMULATION

The One Trial table (above) contains the actual model of the trial - here, for example, the average and maximum of time in
scenario. The headings are the key to expressing the relationships system (i.e., time "on hold" + time being served), time "on hold ,"
between the various model elements as relatively simple (i.e., time server idle and server utilization. [All times are in minutes].
intuitive) formulas. The relationships need only be expressed once, In order to obtain steady state results the above trial has to he
for customer/caller no. I, are then copied down for as many callers repeated a "sullicient" number of times, where each time the
as are required in the simulation window (here 2 hours). Relevant measures of performance are stored in a summary table, shown
measures of performance have to be chosen and evaluated for each below:

Summary of Trials
Time Time Time Time Time Time
in in Server in in Server
System Queue Idle Utilization System Queue Idle Utilization
Avg: 8,9 5,7 0,40 88.4% Max:
18,0 14,5 4.3 98,1%
trial # trial #

7.6 4.0 0.43 89.5% 14.0 10.0 4.0


2 8.6 5.1 0.23 93.9% 2 17.0 15.0 5.0
3 3.8 1.2 1.00 72.4% 3 10.0 7.0 5.0
4 17.3 13.8 0.25 93.4% 4 30.0 27.0 3.0
5 4.8 1.8 0.65 82.3% 5 11.0 6.0 6.0
6 6.0 2.8 0.80 80.0% 6 14.0 9.0 4.0
7 7.8 4.1 0.25 93.6% 7 13.0 11.0 5.0
8 4.7 1.9 0.55 83.5% 8 9.0 7.0 5.0

A straightforward way of determining when the various measures callers "hanging up," multiple servers etc., depending upon the
of performance have stabilized is via graphical presentation. level of the class and the students' familiarity with spreadsheets.
Long-term averages can be used to make recommendations
regarding hiring a second reservation agent. Conclusion
For part (b) essentially the above approach is repeated with The application described above, although not exceedingly
the new arrival distribution. complex, is nevertheless considered to be more rea listic than
It is possible to automate the above procedure by using could normally be attempted without resorting to a simulation
macros, although by no means are they necessary or required. model. Even the least "quantitatively oriented" students persevere
(Students with more advanced spreadsheet experience often elect with implementation and experimentation of these models and
to do this spontaneously.) view the experience as highly beneficial. Two reasons are given
for this: first the use, abuse and relevance of modeling in decision
Discussion making is dramatically demonstrated. And second, the time and
A number of items can be addressed in the discussion of the case: energy students spend in acquiring the appropriate spreadsheet
• Reliability of the recommendations, vis a vis the number of skill levels are perceived by them as acquiring an extremely useful
trials, structure of the model, etc. skill, as an in-depth familiarity with a spreadsheet package is now
• Importance of having relevant and meaningful measures of a necessary requirement for a business major.
performance, e.g., average wait times may be low but if The application described above has been successfully used in
their standard deviation is large the maximum times in line the classroom to demonstrate the principles and usefulness of
may be unacceptable. simulation. Clearly many other (i.e., nonqueuing) situations lend
• Labor standard issues: is a utilization of e.g., 95 percent for themselves to be modeled in the same manner. The case can
a human server appropriate or indeed desirable? readily be updatedlrenewed from semester to semester (1) by
• What are the assumptions in the model, e.g., callers do not changing Alabama Airlines to a bank, movie theater, etc., and
"hang up," changeover times are negligible (or included in (2) by changing the alTival interval and service time distributions.
service times), there are no calls in the system at "time Finally, a not insignificant benefit is the popularity amongst
zero" etc. What are the implicationslrestrictions? students of this approach- where they can dramatically
demonstrate for themselves the consequences to various "what if"
The model and simulation can be easily extended to become more
scenarios by amending a few parameters.
realistic by incorporating the various assumptions and simulating
QUANTITATIVE MODULE F SIMULATION 361

INTERNET CASE STUDIES* Cargo weights:


30 ft. handl es 45 tons
SAIGON TRANSPORT 20 ft. handl es 20 tons
The table in the case represents a cumulative normal distribution It is noted that a truck can carry two 20-foot containers , so
of monthly cargo tonnages. The distribution of cargo between that the total cargo is 40 tons. Thus , ave rage cargo hauled by con­
containerized and noncontainerized cargo is 25%-75%. It is tainerized freight is 53 tons, or 0.6 x 60 + 0.2 x 45 x 0.2 x 0.40.
assumed that the noncontainerizcd shipments will carry the Daily cargo haul ed is : 3 x 53 = 159 tons/day.
capacity cargo of 6 0 tons. Hence, daily fre ight hauled by each The following is simulated monthly freight shipments using
truck is 180 tons (60 tons/trips x 3 trips/day). 30 days/month. Cargo is th en categorized into containerized and
For containerized cargo:
noncontainerized shipments. The daily truck requirements for
60% is packaged in 40 ft. containers
each month is projected.
20% is packaged in 30 ft. containers

20% is packaged in 20 ft. containers

Trucks Trucks
Freight Required Required
per day 3 trips/day 25% 3 trips/day
Mo. RN Freight (30 days) 75% Non-containerized 60 tons/tri p Containerized 53 tons/trip
1 63 171,000 5700 4275 24 1425 9
2 88 197,000 6567 4925 27 1642 10
3 55 165,000 5500 4125 23 1375 9
4 69 176,000 5867 4400 24 1467 9
5 13 124,000 4133 3100 17 1033 7
6 17 131,000 4367 3275 18 1092 7
7 36 150,000 5000 3750 21 1250 8
8 81 186,000 6200 4650 26 1550 10
9 84 190,000 6333 4750 26 1583 10
10 63 172,000 5733 4300 24 1433 9
11 70 177,000 5900 4425 25 1475 9
12 06 110,000 3667 2750 15 917 6

As seen from the simulated year's operation, the daily truck


requirement for noncontainerized cargo ranges from 15 to 27. For
containerized cargo it ranges from 6 to 10 trucks. The maximum
number of trucks required to handl e th e load in any single period
is 37. It should be noted that the utilization factor is 0.96. Hence,
the number of trucks should be adjusted upward accordingly.
Students should simulate future periods with a gi ven Heet
size weighing the opportunity cost of trucks (cost of capital)
against demurrage penalties. Container demurrage should also be
included in the determination.
A discussion of obtaining a simulated "Freight" from a "Ran­
dom Number" should be highlighted; for example, why does the
random number "63" result in a freight of 171 ,OOO?

BIALIS WASTE DISPOSAL, GMBH.


Costs in Euros:
Shipment: Euro 900 per load
Loading/Unloading Euro 120 per load
Overhead (Euro 41 ,000/25) Euro 1,640 pe r load
Euro 2,660 per load
Probabilit;i distributions:
Number of Random Revenue Random
Barrels Number per Number
Loaded Probability Interval Barrel Probability Interval
26-30(28) 0.12
01-12 Euro 50 0.20 01-20
31-35(33) 0.16
13-28 Euro 60 0.44 21-64
36-40(38) 0.24
29-52 Euro 70 0.28 65-92
41 - 45(43) 0.36
53-88 Euro 80 0.08 93-00
46-50(48) 0.12
89-00 1.00
1.00
*Solutions to cases that appear on our companion web site
(www. prenhall.comlheizer).
362 QUANTITATIVE MODULE F SIMULATION

Truckload Random Number of Random Revenue per Total


Simulation Number Barrels Number Barrel Revenue
52 38 06 Euro 50 Euro 1,990
2 37 38 63 60 2,280
3 82 43 57 60 2,580
4 69 43 02 50 2,150
5 98 48 94 80 3,840
6 96 48 52 60 2,880
7 33 38 69 70 2,660
8 50 38 33 60 2,880
9 88 43 32 60 2,580
10 90 48 30 60 2,880
11 50 38 48 60 2,280
12 27 33 88 70 2,310
13 45 38 14 50 1,990
14 81 43 02 50 2,150
15 66 43 83 70 3,010
16 74 43 05 50 2,150
17 30 38 34 60 2,280
18 59 43 55 60 2,580
19 67 43 09 50 2,150
20 60 43 77 70 3,010
21 60 43 08 50 2,150
22 80 43 45 60 2,580
23 53 43 84 70 3,010
24 69 43 84 70 3,010
25 37 38 77 70 2,660
Euro 63,860

A verage income per load = Euro 2,544.40


Loss per load = Euro 2,660 - 2,554.40 = Euro 105.60
Loss per year = 25 x Euro 105.60 = Euro 2,640
The conclusion, based on just one short simulation, is lhal
money will be lost by continuing service to Italy.

You might also like