The intention of this project is to demonstrate the function of
production planning in a non – artificial environment. Through this simulation
we are able to forecast, with a degree of certainty the monthly requirements for
end products, subassemblies, parts and raw materials. We are supplied with
information that we are to base our decisions on. The manufacturer depicted in
this simulation was actually a General Electric facility that produced black and
white television sets Syracuse, New York. Unfortunately this plant is no longer
operational, it was closed down and the equipment was shipped off to China. One
can only wonder if the plant manager would have taken Professor Moily’s class in
production management the plant still might be running.
Modern production management or operation management (OM) systems first
came to prominence in the early half of the twentieth century. Frederick W.
Taylor is considered the father of operations management and is credited in the
development of the following principles.
a. Scientific laws govern how much a worker can produce in a day. b. It is the
function of management to discover and use these laws in operation of productive
systems. c. It is the function of the worker to carry out management’s wishes
Many of today’s method’s of operation management have elements of the
system (MRP) is learning how workers to hire, fire, or lay idle. This is
because it we realize the a worker can only produce so many widgets a day, can
work so many hours a day, and so many days a year.
I will disagree with principle “c” in that the worker should blindly
carry out the wishes of management. Successful operations are based upon a two-
way flow of thought and suggestions from management to labor. This two-way flow
of ideas is incorporated into another modern system of operations management,
the Just – In – Time system. Eastman Kodak gives monetary rewards to employees
who devises an improvement in a current process or suggests an entirely new
process of manufacturing. Often a small suggestion can yield a big reward when
applied to a mass-produced item.
In this project we are presented with the following information: bounds
for pricing decisions, market share determination, the product explosion matrix,
sales history (units per month at average price), unit value, setup man-hours,
running man hours, initial workforce, value of inventory, on hand units. We
also know that we have eight end products, four subassemblies, eight parts, and
four raw materials. The eight end products are comprised entirely from the
subassemblies, parts, and raw materials. From this information I was able to
determine how many units of each final product, how many units of parts to
produce in a month, how many units of raw material to order every month and how
to price the final products.
The first step that I took in this project was to develop product
structures for each product (please refer to the Appendices for product
structures on all eight products, plus new product nine). This information was
presented in product explosion matrix. For example, I determined that product
one used one subassembly nine and one part thirteen. Part thirteen consisted of
raw material twenty-one. Sub-assembly nine consists of part thirteen (which
includes raw material twenty-one), raw material twenty one and raw material
twenty-four. From this product explosion matrix I have realized that an end
product does not just happen; they consist of many subassemblies, parts and raw
We also determined the minimal direct costs to each of the eight
products. The minimal direct product is the cost of the raw material, plus the
price of the amount of labor for the assembly to end product. For product one
we have a total of three raw material “twenty-one” which cost ten dollars a
piece and one raw material “twenty-four” which cost twenty dollars each. We now
have a total of fifty dollars for the price of the parts. Next we calculate the
labor that goes into transforming these parts into a viable end product. We get
a total of six hours of running man hours/unit and an hourly labor rate of $8.50,
which gives us a total of fifty-one dollars. This gives a minimal total cost of
$101 to produce product one. This number is useful in determining how much a
unit actually cost to manufacture and what we must minimally sell the product
for to make a profit. We can than analyze if a product costs to much to make or
the sum of the parts is more than the price of the end product. Product eight
had the lowest direct minimum cost ($89.50) and four had the highest minimal
From a purely economic stand point, it would be beneficial to use as
much of raw material twenty-three ($5 unit) and as little of raw material
twenty-two ($30 unit). This does not consider that raw material twenty-two may
actually be more valuable than raw material twenty-three. Perhaps raw material
twenty two may be gold or silver and raw material twenty-three may be sand or
I also converted all information in the sales history per month (figure
four of the MANMAN packet). The purpose of this step was so that I could sort
and add the sales numbers to chronicle the past twenty four months. Clearly
product one was the best-selling apparatus, and product three, four and five
where sales laggards.
Entering the information into spreadsheet form was also necessary to
present the eight products in graphical form. Of the following graph types that
where at my disposal (line, bar, circle) to clearly illustrate the upward and
downward trend of each of the eight product I chose the line graph method. A
circle graph is good percentage comparisons or comparison of market share. Bar
graphs can illustrate a snapshot in time but can distort trend data.
At this point our class gathered into groups to discuss which product to
discontinue. Obviously product one was not going to be of the discontinued
products, since it was our volume leader. Based on the sales figure for the
past twenty-four months my group decided to eliminate products three, four and
five. Also, products three, four and five had the highest minimum direct costs
as well. Since these products where expensive to manufacture and where our
lowest selling products a group decision was made to discontinue these products.
The discontinued product was then rolled over into a new product, now
referred to product nine. Unfortunately, we where unable to decide by the
information given if any of the discontinued products was a high margin product,
low volume product (IE 50″ big screen color Trinitron tube with oak cabinet and
Moving right into our next step we began to analyze our bar charts to
make our starting forecast. We viewed sales from each product to see if they
fall under one the following situations: Base (Base + Trend) (Base + Trend) *
When a product is base the sales alter little each sales period or change
erratically with external market signals. An example of a product that would
fall under the base model would be sand bags. Sand bags sell at the same level
month after month. If a retailer sells a hundred bags in March the will sell a
hundred bags in October. But, in a flood plain after terrantiel downpour, the
sales of sandbags increase exponentially. This is because many people purchase
the sandbags to hold back the rising flood waters. Another example of a product
that would emulate the base model is insulin. There is a limited number of
people with insulin dependant diabetes. The people with insulin dependant
diabetes unfortunately die off, but are replaced with other people who fall ill
to the disease. There is very little movement up or down in the sale of insulin.
The base plus trend model illustrates that a product has a trend of
upward or downward groth in sales. Products at the begining or ending of their
respective product cycles will display this type of performance. Sales of a new
product such as Microsoft Windows95 disk operating system will fall into this
category. The sales of May are expected to be larger than April, the sales of
April will be larger than March and so on. While the sales may decline (or
increase) during a particular time frame, a trend of upward or downward growth
will be apparent.
Lastly, the base plus trend times seasonality attempts to forecast the
swings in demand that are caused by seasonal changes that can be expected to
repeat themselves during a single or consecutive time period. For example,
florists experience a predictable increases in demand each year, both occur at
similiar (or exact) times during the year; Mothers Day and St. Valentines Day.
Florists must forecast demand for roses and other flowers so they can meet this
predictable demand. If I where to construct a ten year historical graph for a
neighborhood florist, there would be clear increase in demand every February and
May, in every one of those years. A caveat to the previous example would be
that in most lines a business forecasting is never this easy. If it was there
would not be a production management class or operations management science!
Some other methods used to forecast demand are: delphi method,
historical analogy, simple moving average, box Jenkins type, exponential
smoothing and regression analysis. Forecasting falls into four basic types:
qualitative, time series analysis, casual relationships and simulation. All of
the proceeding have pluses, minuses and degrees of accuraccy. I often depends
on the precision of previous data. Also, as is often stated in financial
prospectuses “past performance does not guarantee future results”.
For product one I used base plus trend. The sales started of at 1246
units and gradually increased to 2146 at the end of twenty four months. There
was a slight dip in sales between month nine-teen and month twenty three. This
drop can from internal or external variables.
Product two was little more tricky. The swing where eratic and showed
no detectable trend. I may have been able to use (Base + Trend) * Seasonality
if there was not a decrease in sales from month eight and an increase in sales
in month sixteen. For this I had to employ the base or simle method.
While I find it hard to comprehend how television sales can be seasonal,
products three, five and six fall under (Base + Trend) * Seasonality models. I
was able to replicate the wave in demand with my forecast. Perhaps consumers
are buying portable televisions to use at the beach while on vacation, or people
are replacing there old televisions to watch the Superbowl championship game or
world series. Or maybe even watch the Syracuse Orangemen in the NCAA college
Conceivably, I was reading to much into product six when a decided on
base plus trend model. The way I saw it was that none of the upward or downward
where that substantial when compared with entire data, and sales from month one
(521 units) decreased by almost fifty percent to 242 units.
I felt the same way about product eight that I felt about product two,
this product demostrated eratic swings in no particular trend. I forecasted
demand using the base or simlple method for this product.
From this point I was able to forecast demand. For the safety stock
decision I always tried to error on the side of caution. On average I used a
twenty five percent safety stock level. However, when calculating the MRP or
labor plans I tried to have the minimal amount of surplus. This often means
that I only had safety stock on hand from period to period.
From this project and from the class lectures I have received an
understanding of how how much planning goes into even the most simplest of
manufactured goods. Production managers must employ at least one type of
forecasting method in order to avoid the everyday stock outs, late deliveries
and labor problems that arise. Forecasts are vital to every business
organization and for every significant management decision.
I feel that I could have further reduced costs by reducing some of the
parts, sub assemblies and outsourcing some of the production. Another situation
that I felt was unrealistic was that there was only one source for each part and
when that part was unvailable, there was a stock out. Perhaps in future
projects there can be allowance for this.