You are on page 1of 12

The Australian National University

Assignment 1
SURNAME: Onder
GIVEN NAME: Michal
STUDENT NUMBER: u4798498
SUBJECT: ENGN8005 - Operations Management
WORD COUNT: 3056
u4798498
1

Question 1:

a) Productivity can be expressed as
0nits piouuceu
Laboui houis useu
in our case
boxes
houis
= 3 boxes per labour
hour.

b)

= 3.125 boxes per labour hour.



c) Increase in productivity can be calculated as follows: increased productivity previous
productivity divided by previous productivity. ((3.125-3)/3)*100 = 4.16%.



Question 2:

a) Labour productivity is
iugs
houis
= 0.125 rug per labour hour


b) Simple multifactor productivity is
iugs
houis gallons uays
= 0.102 rug per multifactor.
Dollars can be used as common denominator, thus multifactor productivity would be
iugs

= 65/$8260 = 0.0786 rug per one dollar or 7.86 rugs per 1000 dollars.


Question 3

Oven efficiency or savings can be calculated as difference between last year energy consumption and this
year consumption divided by last year consumption.
(3000-2750)/3000 = 8.33%
It is clear that ovens did not achieve anticipated 15% reduction in energy consumption.

Labour productivity for last year was 1500/350 = 4.29 dozens (51.42) per labour hour
Productivity for this year is 1500/325 = 4.61 dozens (55.38) per labour hour.
Increase in productivity is therefore (55.38-51.42)/51.42 = 7.7%.

For comprehensive understanding of productivity it should be considered other factors such as expenses,
labour, energy consumption. Because there was not denoted dollar value for each factor it is not suitable
to use common denominator such as dollar value, however it can still be used general multifactor
productivity.

1500/(350+1500+3000) = 0.0817 dozens per multifactor
1500/(325+18000+2750) = 0.0712 dozens per multifactor

It can be seen that overall productivity has been decreased. Unless labour and energy costs are
considerably high it would be a decline in productivity even if it is used dollars as common denominator. If
u4798498
2
we consider used energy as a performance measure then there was increase in performance by 9%.
However performance measured by invested money has decreased by 16.6%.

Question 4:
Mission of company can be described in many ways. In general terms it can be expressed as where the
company would like to be and sometimes where it would like to be heading. It can be also depicted as a
statement which defines what a company is, its purpose, who its primary customers are and also
identifies its products and services.
Question 5:
Strategy could be described as plan of actions designed to achieve companys mission. It utilises
opportunities and strengths and eliminates threats and weaknesses. Strategy plan implements mission in
three main conceptual ways:
y Differentiation
y Cost leadership
y Response
Strategy is usually further developed through operations decisions.

Question 6:
Every company has defined mission, which describes where it would like to be and what it would like to
achieve and to achieve it also needs to have a strategy or strategy plan. In other words, mission describes
an organisation and its goals and strategy helps it to get there.

It is clear that there is fundamental difference between mission and strategy. While mission can exist
without strategy, the strategy is based on mission statement and describes action plan linked to mission
goals. Also mission is usually a statement, whereas strategy can be described and developed to long
report.

Question 7:
The mission of our company is to provide high quality service with reasonable (exceptional) price.
We are committed to provide friendly environment to our customers and environment.

The strategy is to achieve high quality service through utilising, cutting edge technology and
equipment together with experienced staff and sophisticated and effective scheduling. We will keep
prices low by precise usage of material and time resources.

Main strategy is based on differentiation and response.
10 OM
Product design
Product is servicing of cars. Cars can be bookend by any electronic means and business hours are
flexible as much as possible. Utilising of garage capacity is main focus to achieve high performance
and low overhead time. This is done by computer scheduling system which can calculate approximate
time needed to finish particular task according to employees input. There is feedback put into
computer after task is accomplished for the future improvement of scheduling system.
u4798498
3
Quality management
Main priority is achieved by using monitoring systems, trained employees and sophisticated
equipment. Quality of services is regularly assessed by statistical means and quality management
plan is developed to identify all faults and keep high quality outcome.

Process design
Servicing of cars is automated and all stations are flexible and modular as much as possible so all
tasks can be performed in any station. All processes are environmentally friendly and monitored to
serve for future improvement.

Location selection
Garage is located in convenient location with easy access by customers, suppliers and employees.
Location should be as close as possible to suppliers and other necessary services to cut additional
costs.

Layout design
Garage is equipped by modern equipment and automated as much as possible. Layout is optimised
to process as much cars as possible.

Human resources
All employees should be highly trained and reliable with clearly defined duties and responsibilities.
There should be hierarchical system from top management to mechanics. All employees are
encouraged to assist with improvement of performance through regular meetings.

Supply chain management
Supplier needs to be flexible, reliable and long term partner with known attributes to allow more
flexible and reliable supply of goods and services to avoid additional expenses and excessive
inventory.

Inventory
Inventory is adaptable and is linked to pre-determined needs for parts and services based on
bookings. There is aim to achieve just in time inventory system to keep storage and other
associated costs as low as possible.

Scheduling
Centralised planning of all booked services is put in place to avoid clashes and to achieve highest
possible capacity. All scheduling is done through central computer. All relevant employees are
trained how to operate it.

Maintenance
All equipment is of high quality with predicted maintenance cost and frequently scheduled for
maintenance. Scheduling system and all operations are monitored and adjusted periodically by
standard managerial means.

u4798498
4
Question 8:
a) Major increase in oil price
This would affect wide spectrum of activities throughout company. It will mainly affect cost of products
and services. Low-cost strategy would need to be further re-developed to keep costs low. Also companies
which rely on transport and oil products would need to adjust their strategies.
It would mainly affect product design, process design, supply chain management, inventory scheduling
and maintenance. Product design would need to be more energy efficient and less depended on oil
products and likewise for process. Supply chain management would need to be more efficient and also
there should be considered different ways of transport such as train. Scheduling would need similar
changes as supply chain management together.

b) Water and air quality legislation
This would affect product design, quality management, process design, and location selection supply
chain. Design of product would need to utilise environmentally friendly materials and processes. This is
directly linked to higher costs so design needs to be highly efficient. Quality management would need to
be adjusted to meet new legislations requirements. Location selection might help to get around new
requirements by outsourcing of critical processes to area with more benevolent legislations. The same
would need to be applied to supply chain.

c) Fewer prospective young employees entering the labour market
This is increasingly issue in developed economies where fewer young prospective people enter the labour
market. This would affect mainly human resources strategy, where the emphasis would be given to retain
prospective employees. Also there would need to be strong strategy for developing and training of
existing employees to achieve high productivity without hiring of new people with required knowledge.
There would need to be strategy for social activities and benefits for existing employees to secure them in
long therm. When there is weak supply on labour market it is usually more expensive to get new an
employee than to retain existing. There could be strategy to hire employees from overseas or use
outsourcing of labour to countries with higher amount of prospective young employees.
d) Inflation versus stable prices
This would affect whole range of aspects of company strategy. Usually, stable prices give opportunity to
undertake cost expensive activities such as major design development, building of new premises acquiring
of materials and services, opening of new branches and so on. In other words it is time for growing and
developing. Inflation has exactly opposite effect; however it also offers an opportunity to get better
position in future if there is ability to secure capital and resources to do that. For instance, companies in
strong and stable financial situation can invest and develop even during inflation and get advantage in
future in compare with other companies.

e) Legislation moving health insurance from a benefit to taxable income
This would not directly affect company strategy other than human resources strategy and companies
which has a main asset their employees. This would require adjustments in salary packaging through
either rising of salary package to compensate tax loss or by utilising other benefits which are not taxable.
Should the legislation be very adverse, there would be also an opportunity to outsource employees to
region with more friendly legislation.



u4798498
5
Question 9:

Time series model uses assumption that future values are function of previous values from past and
present. It basically uses existing data from past to predict future based on behaviour from past. For
example to predict sale of cars per next month it could be used data from previous month to
calculate demand. Also, time series model shifts through existing to recent data as it is available. It is
continuing process. It is usually not expressed as explicit formula. It is not possible to calculate it once
and use without constantly add new data.

Causal or associative model also uses previous data but it does not relay purely on it to predict
future in continuous way as time series model. In other words, it uses existing data to generate
explicit equation which can be used then independently from existing data. It can also incorporate
variables and factors into equation that can influence forecast. That does not mean it cannot be
adjusted if it is required.

Question 10:

Smoothing constant () is typically chosen in range from 0.05 to 0.5. If it is chosen higher than 0.5
and close 1.0, the forecast model is equal to naive model where F
t
= A
t 1
.
From equation,
F
t
= F
t -1
+ (A
t-1
F
t-1
)
is clear that higher value increases weight give to recent value. Where = 1 weight reaches also 1
and previous forecast is eliminated. When is closer to zero, exponential smoothing model gives
higher value to past data.
Question 11:
Seasonal indexes correspond with seasonal variations in data. Seasonal variations are regular up and
down movements due to specific events such as weather, peak hours, sales time of day and so on.
Value of seasonal index represents by how much (many) is value bigger or smaller in compare with
average value for seasonal cycle (hour, day, year, etc). For instance, if seasonal index is 1.05, it would
mean that value for this particular period of time is 5% larger than average value for whole seasonal
period.
Seasonal patterns occur regularly from several hours to several weeks or months. This is due to
variations of external variables such as weather, business hours, sales, sport events and others. It is
easy to find correlating variables.

Cyclical patterns occur every few years or even decades. It is considerably harder to identify them
mainly because of missing obvious correlating variables. However once the variable is identified it is
possible to apply standard statistical methods to analyse and predict these patterns.




u4798498
6
Question 12:
Exponential smoothing model uses following formula:
F
t
= F
t -1
+ (A
t-1
F
t-1
)
Where is smoothing coefficient and F
0
is initial forecast. In our case =0.5 and initial forecast is 30.
In order to control forecast error is vital to utilise some form of measure. Popular method is MAD
(mean absolute deviation) and also tracking signal.
MAD is calculated as follows,


Tracking signal is calculated as


Where RSFE (running sum of the forecast errors) is calculated as



Fig 1, Sales of music stands and forecast.
MAD should be as low as possible and there is desire to try several smoothing coefficients to
determine lowest MAD.
Tracking signal should be in predefined range. It is not recommended use range higher than 8
MADs. If the tracking signal is higher or lower than this value, it indicates problem with forecasting
method or some error in calculation or input data error.

Fig 2, Tracking signal of forecast error.
0
5
10
15
20
25
30
35
40
0 2 4 6 8 10 12
actual demand
forecast
-3.0
-2.0
-1.0
0.0
1.0
2.0
0 2 4 6 8 10 12
M
A
D
Tracking signal
u4798498
7
Question 13:
a) Ridership for summer months in London subway is plotted below;

Fig 3, Ridership in London subway.
It can be seen linear relationship between number of tourists and sold tickets in subway.
b) In order to be able to analyse and predict future ridership it can be used least square method
to obtain linear regression model. It can be expressed by following equation,
= a + bx
where constant a and b are calculated from given data.
a = and b=

where x represents number of tourists and y number of


tickets or ridership.
n = 12
132/12 = 11
27.1/12 = 2.26
xy = 352.9

= 1796
b = (352.9 - 12*11*2.26)/(1796 12*11*11) = 0.159
a = 2.26 0.159*11 = 0.506
thus, = 0.506 + 0.159x

Detailed calculation and table of data is in appendix.

c) Ridership for 10 million visitors can be calculated by substituting 10 into given equation,
Y = 0.506 + 0.159*10 = 2.096 or 2.1 million.

y = 0.1593x + 0.506
R = 0.8403
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 5 10 15 20 25
Ridership
Linear (Ridership)
u4798498
8
d) If there are no tourist, the ridership can be calculated by adding 0 to equation, thus
y = 0.506
This represents hypothetical ridership only by residents of London. The relevancy of this
number would be hard to determine as the function of ridership can be distorted when
number of tourists is low. It is only safe to use obtain equation in a range between 2 million
and slightly beyond 20 million tourists.
e) Standard error of estimate represents accuracy of regression estimate. It represents error
from depended variable (ridership) to regression line. In other words, this value can be seen
as a range for estimated value. It can be calculated as:
S
y,x
=

= 0.407
f) Correlation coefficient (r or R) expresses the degree of strength of linear relationship. It can
also be calculated coefficient of determination R
2
. This represents how many percent of
variable y is explained by linear regression.
r =

= 0.917
r
2
= 0.840
It is clear that there is strong positive linear relationship and around 84% of data is explained
by linear regression.

Question 14:
Dear John,

I reviewed and analysed given data and applied standard managerial tools to assess current situation and
to develop tools for better planning and prediction of future demands.

Firstly, I analysed each year separately to identify any seasonal patterns. From figure below it is clear that
every year there were several peaks and troughs. This should help us in better prediction of future
demands month by month.


Fig 4, Historic number of orders of phone cases in 2000-2002.
December, January & July peak times
February, April & August troughs
400
450
500
550
600
650
700
0 1 2 3 4 5 6 7 8 9 10 11 12 13
2000
2001
2002
u4798498
9
Secondly, I analysed average demand over three years. It can be seen that apart from seasonal variations
overall demand was constantly growing in all three years. There is clear trend over the period of time.
This will help us to predict overall demand in next year as well as in years to come.

Thirdly, I have analysed seasonal patterns and trends to develop model for prediction of demands month
by month in coming year. This model is accurate just over couple of months. However it is possible to
adjust it every month with actual demand to get a new prediction.

Main report:

Average demand for next year can be calculated by averaging of previous years and by analysing of this
data in order to find linear relationship. It is clear from graph that this relationship is almost exact line. I
have applied a linear regression (method of least squares) to obtain formula for prediction of next year
demand.


Fig 5, Average number of orders of phone cases in 2000-2002.
Average demand for 2003 = a + bx = 401.36+4*68.292 = 675 cases.
Please note that number 4 represents year 2003.

This is however average demand for whole year and does not consider seasonal demands. We could use
this for prediction for overall planning of resources.

To be able to predict demand month by month I have utilised seasonal time series forecast method which
consists of trend projection and seasonal adjustment for each month. This is called multiplicative seasonal
model. This method uses seasonal indexes which are calculated from past data per each month.
The way how it works is following.
y Firstly we calculate average demand for each year.
y Secondly we calculate average demand for each month from past data for all years.
y Thirdly we divide each average month demand by overall average demand.
This will create index per each month. This index represents by how much is demand different from
average whole year demand. In next step we will multiply already computed predicted demand for next
year by seasonal indexes to obtain prediction of number of orders per each month.
y = 68.292x + 401.36
R = 0.9939
0
100
200
300
400
500
600
700
0 1 2 3 4
Number of orders
Linear (Number of
orders)
u4798498
10
I have also used another technique called exponential smoothing with trend adjustment for comparison
and proofing of the first method. This method can also be used for prediction of future demands.


Fig 6, Orders of phone cases in 2000-2002 and predicted demand in 2003.
Horizontal axis denotes to number of months from January 2000 to December 2003

I have also created tool for analysing of forecasting model. This tool is using MAD (Mean Absolute
Deviation) and tracking signal to assess relevancy of model. Tracking signal should be in range 2 MADs,
however it is acceptable for variation up to 8 MAD.
This will help us to control quality of prediction model and make adjustments accordingly.


Fig 7, Tracking signal of forecast error.

Detailed calculations and tables of data are attached in appendix.

I believe this report give you comprehensive view of current situation and future predictions.
Should you have any query do not hesitate to discuss this matter after holidays.





y = 5.2484x + 440.85
R = 0.7621
300
350
400
450
500
550
600
650
700
750
800
0 5 10 15 20 25 30 35 40 45 50 55
Past demands
Predicted demands
forecast trend
forecast demand
Linear (Past demands)
-10.0
-5.0
0.0
5.0
10.0
0 5 10 15 20 25 30 35 40
M
A
D
Tracking signal
u4798498
11
References:

[1] Heizer J., Render B., Operations Management, 8
th
revised edition, Pearson education, 2007.
[2] Warrack B., Keller G., Statistics for management and economics, 6
th
revised edition, Thomson,
2003.

You might also like