Chapter         3

Forecasting

3.1 Introduction
Forecasting is a method to use the past experience and estimate the future. This is a very
critical planning tool. It can be used for sales forecasting, demand forecasting and technology
forecasting. In production engineering, demand estimation is an important part of production
planning. There are two types of demands: dependent and independent demands. The quantity of
dependent demand is estimated by the demand of complete or end product. But for the estimation
of independent demand, various qualitative and quantitative models of forecasting are used.
Forecasting is the initial phase of production planning in which the quantity of product required
in the near future is estimated and production schedule is prepared accordingly. The term ‘forecasting’
can be defined in many ways. Some of the definitions given by the researchers are as follows:
‘Forecasting is predicting, projecting, or estimating some future event and condition
which is outside the organization’s control and provide for basis of managerial planning’.
(Golden et al. 1994)
‘Forecasting is generally used to predict or describe what will happen given a set of
circumstances or assumptions’. (Waddell 1994)
‘Forecasting is a projection into the future of expected demand, given a standard set of
environmental conditions’. (Mentzer and Moon 2005)
In the long-term forecasting, the demand of the product, change in design in the product
(technology forecasting) and demand of new product are estimated; and expansion of plant
or requirement of additional resources is estimated accordingly. Thus, forecasting is a tool to
estimate quantitative changing in demand in future. This may be based on trends of past data on
actual demand or experts’ opinion depends upon the type of forecasting.

3.1.1 Characteristics of Forecasts

Forecasts have the following characteristics.
1. Forecasts are normally wrong since it is an approximation of demand estimates in the
future.
2. A good forecast has more than a single number, as it includes
(a) A mean value and standard deviation.
(b) An accuracy range (high and low).
 Forecasting  51

3. Aggregated forecasts are usually more accurate because it can adjust the variation in
actual demand of individual products.
4. Accuracy erodes as we go further into the future because the increase in time horizon
increases the uncertainty in demand pattern.
5. Forecasts should not be used to the exclusion of known information.

3.1.2 Forecasting Horizons

Forecasting can be classified in terms of time span they will cover in the future. The basic types
of time horizon forecasts are long term, medium term and short term (Korpela and Tuominen
1996). The ranges of these three forecasting horizons are: long-range forecasting (for more than
2 years), medium-range forecasting (for 3 months to 2 years) and short-range forecasting (for
less than 3 months). Long-range forecasting is normally used for higher investments such as
expansion factory capacity, location of the new plant/facilities, launching of the new product,
etc. Medium-range forecasting is used for production planning, sales planning, increasing
department capacities, etc. Short-range forecasting is used to schedule the production, work
assignment, sequencing, etc. The time span and examples of these forecasting horizons are
described in Table 3.1.

Table 3-1: Forecasting horizons and their examples

Forecast horizon Time span Item being forecast

• Product lines
• Factory capacities
Long range • Planning for new products
Years
(More than 2 years) • Capital expenditures
• Facility location or expansion
• R&D
• Product mix
Medium range
• Department capacities
(3 months to 2 years) Months
• Sales planning
• Production planning and budgeting
• Specific product quantities
• Machine capacities
• Planning
Short range • Purchasing
Weeks
(Less than 3 months) • Scheduling
• Workforce levels
• Production levels
• Job assignments

3.1.3 Steps of Forecasting

To forecast demand, the following essential steps need to be taken:
1. Determination of objectives of forecasting: The objectives of forecasting should be
clear. Without focused objective, forecasting is meaningless.
52  Industrial Engineering and Management

2. Selection of the items to be forecast: The forecast for one item cannot be used for the
other item. Therefore, the item must be selected for which forecast is to be made.
3. Determination of the time horizon of the forecast: The nature of the product decides
the time horizon of its demand forecast. A long-range forecast is more error-prone as
compared to a short-range forecast. Therefore, the horizon or a range of forecasting is
an important issue.
4. Selection of the suitable forecasting model: There are a number of forecasting models.
The suitability of these models is to be checked for the forecast of a particular product.
The choice should be of a forecasting model that would give a more accurate estimation
of demand.
5. Collection of the data: Future is predicted on the basis of the past data. Therefore, the
relevant past data should be collected from various sources such as annual reports,
magazines, journals, etc.
6. Validation of the forecasting model: The selected forecasting model is to be validated
using the collected data.
7. Forecasting the demand: After data collection and model validation, forecast is made
for a specific time in the future.
8. Implimentation of the results: Finally, the results are implemented to fulfil the objectives.
Here, results mean the quantity/volume of product forecasting. The entire production
systems are changed/adjusted to meet the expected demand in future.
When carrying out demand forecasts, one often confronts with the problem of the inappropriate
selection of a forecast method. It should be noted that in every actual forecast situation
methods have their advantages and disadvantages; hence, it is important to define and analyse
forecast method selection criteria (Pilinkiene 2008). In order to select the appropriate method,
several criteria should be considered such as the degree of forecast accuracy, time span, the
amount of necessary initial data, forecast costs, result implementation and applicability level
(Pilinkiene 2008).

3.2 Forecasting Methods

Forecasting methods are divided into two categories: qualitative forecasting and quantitative
forecasting.

3.2.1 Qualitative Forecasting

Qualitative forecasting is usually based on judgements about causal factors that underlie the
demand of particular products or services. Qualitative forecasting can be used to formulate
forecasts for new products for which there are no historical data. Qualitative techniques
provide the means to adjust the forecasts using the experience and judgement of people
knowledgeable about the product being forecast and the environment affecting the forecast.
In other words, one can say that qualitative forecasting emphasizes predicting the future.
Some qualitative methods of forecasting are described below as:
 Forecasting  53

1. Executive committee consensus: This method of qualitative forecasting is based on the

consensus of the committee of executives formed for making forecasts about specific
items. Senior managers draw upon their collective wisdom to map out future events.
They conduct discussions in open meetings. However, these discussions may subject to
the drawbacks of group think and personality dominance.
2. Delphi method: In the Delphi method, at least two rounds of forecasts are obtained
independently from a small group of experts. This group can be between 5 and 20
experienced and suitable experts and poll them for their forecasts and reasons
(Armstrong et al. 2005). The experts never actually meet and typically do not know who
the other panel members are (Wisniewski 2006). After each round, the experts’ forecasts
summed up and reported back to the experts. The cycle can go on from a second to a
third round and so on if appropriate (Lancaster and Lomas 1985). Generally the Delphi
method is used to produce a narrow range of forecasts rather than a single view of the
future (Wisniewski 2006).
In this method, the forecasts of a panel of experts are elicited through a sequence of
questionnaires. Each questionnaire builds on the previous questionnaire in a series of
‘rounds.’ The experts answer the questionnaires independent of each other. After each
round, a facilitator provides an anonymous summary of the experts’ forecasts from the
previous round as well as the reasons they provided for their judgements. The experts are
encouraged to revise their earlier answers in the light of the replies of other members of
their panel. It is believed that the group will eventually reach a consensus. The process
is stopped after a number of rounds or when a consensus is reached.
3. Survey of sales force: The survey of the sales force method is a sales forecasting
technique. It predicts future sales by analysing the opinions of salespeople as a group.
The sales people continually interact with customers, and from this interaction they
usually develop a knack for predicting future sales. The survey of the sales force
method is considered a valuable management tool and is commonly used in business
and industry throughout the world. This method can be further improved by providing
the sales people with sufficient time to forecast and by offering incentives for accurate
forecasts. Companies can make their salespeople better forecasters by training them to
interpret effectively their interactions with customers.
4. Expert evaluation technique: Expert evaluations are made using the experience of
people, such as executives, salespeople, marketing people, distributors or outside
experts, who are familiar with a product line or a group of products, to generate sales
forecasts. The advantage of soliciting contributions from more than one person for
making expert evaluations is that it can offset biases introduced into a forecast that
produced by one person.

Advantages of Qualitative Forecasting

There are following advantages of qualitative forecasting:
1. Qualitative forecasting techniques have the ability to predict changes in sales patterns.
2. Qualitative forecasting techniques allow decision-makers to incorporate rich data
sources consisting of their intuition, experience and expert judgement.
54  Industrial Engineering and Management

Disadvantages of Qualitative Forecasting

There are following disadvantages of qualitative forecasting:
1. In a qualitative method, the ability to forecast accurately suffers due to the following
reasons:
(a) Forecasters only consider readily available and/or recently perceived information.
(b) The forecasters are unable to process large amounts of complex information.
(c) The forecasters are overconfident about their ability to forecast accurately.
(d)  Political factors within organizations as well as political factors between
organizations may influence the forecasting accuracy.
(e) The forecasters tend to infer relationships or patterns in a data when there are no
patterns.
(f) Anchoring on forecasts implies that forecasters get influenced by initial forecasts
(e.g. those generated by quantitative methods) when making qualitative forecasts.
2. The future ability to forecast accurately may be reduced when a forecaster tries to
justify, rather than understand a forecast that proves inaccurate information.
3. Qualitative forecasting techniques lead to inconsistencies in judgement due to moods
and/or emotions of forecasters as well as due to the repetitive decision-making inherent
in generating multiple individual product forecasts.
4. Qualitative forecasting techniques are expensive and time-intensive.

3.2.2 Quantitative Forecasting

Quantitative forecasting is based on the assumption that the ‘forces’ that generated the past
demand will also generate the future demand, i.e., history will tend to repeat itself. Analysis
of the past demand pattern provides a good basis for forecasting the future demand. Majority
of quantitative approaches fall in the category of time-series analysis. Quantitative methods of
forecasting are divided into two categories: (a) time-series methods and (b) causal methods.

3.3 Time-series Forecasting

The time-series forecasting methods are based on the analysis of historical data. Time series may
be defined as a set of observations measured at successive times or over successive periods. In the
time-series methods, the assumption is that the past patterns in the data can be used to forecast
future data points. Here, future data point means the points on the projection of the line using
extrapolation. The patterns of demand variations of a time series are shown in Figure 3.1.
Trends are noted by an upward or downward sloping line. There is a linear relationship
between time and demand. The demand in the future can be estimated by extending the straight
line or extrapolation. The linear trend is shown in Figure 3.1(a).
Seasonality is a data pattern that repeats itself over the period of one year or less. It leads rise
or fall in demand among the seasons. In a particular season, demand may increase while in other
season demand may decrease. Thus, the demand is dependent on season or quarter of a year. The
seasonal variation in demand is shown in Figure 3.1(b).
 Forecasting  55

Seasonal pattern
with linear
growth
Demand

Demand
Increasing
linear trend

Time Time

(a) (b)
Purely random-
No recognizable
Cyclic pattern
pattern
Demand

Demand

Time Time

(c) (d)

Figure 3-1: (a) Linear variation, (b) seasonal variation, (c) cyclic variation and (d) random variation

Cycle is a data pattern that may cover several years before it repeats itself. Cyclic nature in
demand variation shows the same style in increasing or decreasing in demand in future. If we
draw a linear line through the mean values of the actual demands, we can observe that the upper
and lower variation in demands follow the same pattern throughout the straight line as shown in
Figure 3.1(c).
Random fluctuation (noise) results from random variation or unexplained causes. There are
a number of causes of random fluctuation in demand, for example, anticipation of increasing
the price or shortages of product in near future increases the demand. Random fluctuations in
demand are shown in Figure 3.1(d).
The following methods are for time-series forecasting:
(a) Naïve method
(b) Simple moving average (SMA) method
(c) Weighted moving average (WMA) method
(d) Exponential smoothing method

(A) Naive Method

The naive forecasting model is a special case of the moving average forecasting model where
the number of periods used for smoothing is 1. Therefore, in the naive method, the forecast for a
period, t, is simply the observed value of the previous period, t – 1. Due to the simplistic nature
56  Industrial Engineering and Management

of the naive forecasting model, it can only be used to forecast up to one period in the future. It is
not at all useful as a medium-/long-range forecasting tool.
The naive method is used partly for completeness and partly for its simplicity. It is unlikely
that any one will want to use this model directly because it can be only used for the just next
period and also, the forecast for the next period is same as the previous actual demand, i.e. there
is no change is considered in forecasting the demand. Instead, consider using either the moving
average model or the more general WMA model with a higher (i.e. greater than 1) number of
periods, and possibly a different set of weights.
(B) Simple Moving Average Method
The SMA method uses the average value of actual demand for some recent periods. The value of
averaging period depends on the nature of variation in the actual demands for last some periods
and accuracy of forecasting. Using this method, we can forecast the demand only for the next
period in the future, but to know the forecasting errors we have to find the forecast value based
on the actual demands for the past periods, also. Suppose t represents the current period and we
want to forecast for the period t + 1. Specifically, the forecast for period t + 1 can be calculated
at the end of period t (after the actual demand for period t is known) as
1
Ft +1 = ( Dt + Dt −1 +  + Dt +1− n )
n
1 t
⇒ Ft +1 = ∑ Di
n i = t +1− n
where D indicates the demand and F indicates the forecast; t is the time period; n is the number
of the averaging period.
The assumptions in making the forecast using the SMA method are as follows:
1. All n past observations are treated equally.
2. Observations older than n are not included at all.
3. It requires that n past observations be retained.

Example 3.1: The monthly demands for an office furniture (in units) are given in Table 3.2.
Forecast the demand using 3-period and 5-period SMA for the 12 th month. Also, show the
variation in forecasting graphically.

Table 3-2: Monthly demands for furniture for eleven months

Month (t) 1 2 3 4 5 6 7 8 9 10 11 12
Demand (Dt) 600 628 670 735 809 870 800 708 842 870 739 —

Solution:
We use the formula
1 1 t
Ft +1 =
n
( Dt + Dt −1 +  + Dt +1− n ) = ∑ Di
n i = t +1− n
 Forecasting  57

For 3-period moving average method

Here, the average value of last three periods is used as forecast for the fourth period as shown in
the formula below:

1 1
F4 = ( D1 + D2 + D3 ) = (600 + 628 + 670) = 632.66 ≈ 633
3 3
Similarly, F5, F6, …, F12 can be calculated as:
1
F5 = ( D2 + D3 + D4 ) = 678
3
1
F6 = ( D3 + D4 + D5 ) = 738
3
……………………..
1
F12 = ( D9 + D10 + D11) = 817
3
For 5-period moving average method
Here, the average value of last five periods is used as forecast for the sixth period as shown in
the formula below:
1
F6 = ( D1 + D2 + D3 + D4 + D5 )
5
1
= (600 + 628 + 670 + 735 + 809) = 688.4 ≈ 689
5
Similarly, F7, F8, …, F12 can be calculated as:
1
F7 = ( D2 + D3 + D4 + D5 + D6 ) = 743
5
1
F8 = ( D3 + D4 + D5 + D6 + D7 ) = 777
5
…………………………………
1
F12 = ( D7 + D8 + D9 + D10 + D11 ) = 792
5
The forecasting of the demand for the 12 months is shown in Table 3.3 using 3-period simple
average and 5-period SMA.

Table 3-3: Forecasting of furniture demand using simple moving average (SMA)
Month Demand 3-Period moving average 5-Period moving average
1 600 — —
2 628 — —
(Continued )
58  Industrial Engineering and Management

Table 3-3: Forecasting of furniture demand using simple moving average (Continued)
Month Demand 3-Period moving average 5-Period moving average
3 670 — —
4 735 633 —
5 809 678 —
6 870 738 689
7 800 805 743
8 708 827 777
9 842 793 785
10 870 784 806
11 739 807 818
12 – 817 792

It can be observed from the graph in Figure 3.2 that the variation in actual demand is very
high and it is difficult to forecast using any trend line. Thus, the moving average method is
suitable for this kind of data. The 5-period moving average graph is smoother than the 3-period
moving average graph, but the limitation is forecasting error. A larger periods of averaging show
large variation from the actual demand curve as shown in Figure 3.2. Three-period averaging is
closer the actual demand capered to five-period averaging.
900

850
Demands (in units)

800

750
Demand
700 3-Period moving average
650 5-Period moving average

600
1 2 3 4 5 6 7 8 9 10 11 12
Months

Figure 3-2: Variation in furniture demand with time

(C) Weighted Moving Average Method

The WMA method is very similar to SMA method, but in the former method different weights
are provided for the periods. The largest weight is provided with most recent period and the
weights are decreasing as we move to the previous period in the past. A WMA is a moving
average where each historical demand may be weighted as:
Ft +1 = W1 Dt + W2 Dt −1 +  + Wn Dt +1− n
 Forecasting  59

where n is the total number of periods in the average, Wt is the weight applied to period t’s
demand, W1 > W2 >…> Wn, sum of all the weights = 1, Forecast Ft + 1 = forecast for period t + 1.
The assumptions in making the forecast using the SMA method are as follows:
1. Adjustments in the moving average to more closely reflect fluctuations in the data.
2. Weights are assigned to the most recent data.
3. Requires some trial and error to determine precise weights.

Example 3.2: Using the data shown in Table 3.3, forecast the demand for the periods with the
three-period WMA method. The most recent data should be given 50 per cent weightage, second
year past data, 30 per cent, and third year past data, 20 per cent. Also, compare the result with
the result of three-period SMA graphically.
Solution:
Ft +1 = W1 Dt + W2 Dt −1 +  + Wn Dt +1− n
F4 = W1 D3 + W2 D2 + W3 D1 = 0.5 × 670 + 0.3 × 628 + 0.2 × 600 = 643.4 ≈ 644

Similarly, F5, F6, …, F12 can be calculated as:

F5 = W1 D4 + W2 D3 + W3 D2 = 695

F6 = W1 D5 + W2 D4 + W3 D3 = 759
………………………….
F12 = W1 D11 + W2 D10 + W3 D9 = 799
The forecasting of the demand for the 12th month is shown in Table 3.4.

Table 3-4: Forecasting of furniture demand using weighted

moving average (WMA) method
Month Demand 3-Period WMA
1 600 —
2 628 —
3 670 —
4 735 644
5 809 695
6 870 759
7 800 825
8 708 823
9 842 768
10 870 794
11 739 830
12 — 799
60  Industrial Engineering and Management

900

850
Demands (in units)

800
Demand
750

700 3-Period moving

average
650 3-Period weighted
moving average
600
1 2 3 4 5 6 7 8 9 10 11 12
Months

Figure 3-3: Variation in furniture demand using 3-period SMA and WMA methods

In Figure 3.3, we observe that the WMA forecast closer to the actual demand compared to
the SMA method, i.e. the forecasts from WMA pursuing the actual demand closely since the
recent period is given more weightage.
(D) Exponential Smoothing Method
Exponential smoothing is the most popular and cost effective of the statistical methods. It is based
on the principle that the latest data should be weighted more heavily and ‘smoothers’ out cyclical
variations to forecast the trend (Armstrong et al. 2005). It assumes that the data gets older,
becomes less relevant and should be given less weight. In order to calculate the forecasting, the
old average, the actual new demand and a weighting factor are needed.
Exponential smoothing gives greater weight to demand in more recent periods, and less
weight to demand in earlier periods as shown in Figure 3.4. It is a sophisticated WMA method
that calculates the average of a time series by giving recent demands more weight than earlier
demands. In this method, both the actual demands and forecast of the past demands are used in
the calculation of forecasting, i.e. all the past data are used to estimate the demand in the future.
Here, α is a smoothening factor, its value is decreasing as we move towards the past. Figure 3.4
shows the decreasing pattern for the value of α as we move towards the past; this decrease
follows an exponential pattern. Thus, this method is known as exponential smoothing method.
0<α<1

Decreasing weight given α

to older observations
α (1 – α)
α (1 – α)2

α (1 – α)3

Present

Figure 3-4: Decreasing weight to the older observations

 Forecasting  61

The forecast for the period t + 1 is equal to the actual demand for the period t plus the α times of
the difference of the actual and forecasting value for the period t. Here, α tries to smoothen the
variation in the previous period actual and forecasting values of the demand.
The forecast for tth period can be given as

Ft +1 = α Dt + (1 − α ) Ft = Ft + α ( Dt − Ft )

Here, α is smoothening factor.

Example 3.3: Using the data shown in Table 3.3, forecast the demand for the periods using the
exponential smoothing method (α = 0.3 and α = 0.5). Also, compare the results graphically.

Solution:
Using the formula

Ft +1 = α Dt + (1 − α ) Ft = Ft + α ( Dt − Ft )

The demand forecast data is prepared (see Table 3.5).

Table 3-5: Furniture demand forecasts using exponential smoothing

Month Demand Forecast, Ft
α = 0.3 α = 0.5
1 600 — —
2 628 600 + 0.3(600 – 600) = 600 600
3 670 600 + 0.3(628 – 600) = 608.4 614
4 735 608.4 + 0.3(670 – 608.4) = 626.8 642
5 809 626.8 + 0.3(735 – 626.8) = 659.3 688.5
6 870 659.3 + 0.3(809 – 659.3) = 704.2 748.7
7 800 704.2 + 0.3(870 – 704.2) = 753.9 809.3
8 708 753.9 + 0.3(800 – 753.9) = 767.7 804.6
9 842 767.7 + 0.3(708 – 767.7) = 749.8 756.3
10 870 749.8 + 0.3(842 – 749.8) = 777.4 799.1
11 739 777.4 + 0.3(870 – 777.4) = 805.2 834.5
12 – 805.2 + 0.3(739 – 805.2) = 785.3 786.2

For smaller values of a, we get a smoother curve. In Figure 3.5, it can be observed that the
curve for a = 0.3, is smoother than that of the curve for a = 0.5. For smooth curve, forecasting
is easy, but the forecasting error may be large because there is a large deviation of the forecast of
the actual value. This type of forecasting model is used for the demand fluctuates continuously
and there is requirement of smoothening the curve for ease of forecasting.
62  Industrial Engineering and Management

900
850
Demands (in units)

800
750
Demand
700 α = 0.3
650 α = 0.5
600
1 2 3 4 5 6 7 8 9 10 11 12
Months

Figure 3-5: Variation in forecasting using exponential smoothing factor 0.3 and 0.5

Features of Exponential Smoothing Method

Some of the features of exponential smoothing method are enumerated as follows.
1. The emphasis given to the most recent demand levels can be adjusted by changing the
smoothing parameter.
2. Exponential smoothing is simple and requires minimal data.
3. Larger α values emphasize recent levels of demand and result in forecasts more
responsive to changes in the underlying average.
4. Smaller α value treats past demand more uniformly and result in more stable forecasts.
5. When the underlying average is changing, results will lag actual changes.
6. The new forecast is the weighted sum of old forecast and actual demand.
7. Only two values, i.e. actual demand and forecast for just previous period (Dt and Ft), are
required, compared with actual demands for n period in moving average method.
8. Parameter α is determined empirically (whatever works best). But, the rule of thumb
can be used as: α < 0.5.
9. Typically, α = 0.2 or α = 0.3 works well.

Adjusted Exponential Smoothing

Double exponential smoothing is also called exponential smoothing with trend. If trend exists,
single exponential smoothing may need adjustment. There is a need to add a second smoothing
constant to account for the trend. It is similar to a single exponential smoothing. The basic idea
is to introduce a trend estimator that changes over time. If the underlying trend changes, over-
shoots may happen. Issues to choose two smoothing rates, α and b are very important.
The following points may help in choosing the value of a and b:
• b close to 1 means quicker responses to trend changes, but may over-respond to random
fluctuations.
• α close to 1 means quicker responses to level changes, but again may over-respond to
random fluctuations.
 Forecasting  63

Adjusted forecasting is given by

AFt +1 = Ft +1 + Tt +1

Here, T is an exponentially smoothed trend factor given by

Tt +1 = β ( Ft +1 − Ft ) + (1 − β )Tt

where Tt is the last period trend factor and β is a smoothing constant for trend. The formula for
the trend factor reflects a weighted measure of the increase or decrease between the next forecast,
Ft + 1, and the current forecast, Ft. β is a value between 0.0 and 1.0 and reflects the weight given to
the most recent trend data. It is usually determined and subjectively based on the judgement of
the forecasters. High β reflects that trend changes more than in the case of low β. It is common
for β to equal α in this method.

Example 3.4: Using the data shown in Table 3.6, forecast the demand for the periods with
adjusted exponential smoothing method ( α = 0.5 and β = 0.3). Also, compare the result
graphically with simple exponential smoothing (α = 0.5).

Solution:
Adjusted forecast for period 3:
We have

T3 = β ( F3 − F2 ) + (1 − β )T2
= ( 0.3) ( 614 − 600 ) + ( 0.7 )( 0 ) = 4.2
Therefore,
AF3 = F3 + T3 = 614 + 4.2 = 618.2

Table 3-6: Adjusted exponential smoothing forecasts

Month 1 2 3 4 5 6 7 8 9 10 11 12
Demand 600 628 670 735 809 870 800 708 842 870 739 –
α = 0.5 600 614 642 688.5 748.7 809.3 804.6 756.3 799.1 834.5 786.7
AFt 600 618.2 653.34 710.38 782.07 850.84 832.26 761.17 815.34 856.48 787.7
Tt 0 4.2 11.34 21.88 33.37 41.54 27.66 4.87 16.24 21.98 1.04

Single exponential smoothing forecasting with α = 0.5, adjusted exponential smoothing

(double exponential smoothing), and actual demand are shown in Figure 3.6. AFt shows the
curve with trend adjustment and Ft shows the curve for exponential smoothing without trend
adjustment.
64  Industrial Engineering and Management

900

850
Demands (in units)

800

750 Demand
700 α = 0.5

650 AFt

600
1 2 3 4 5 6 7 8 9 10 11 12
Months

Figure 3-6: Variation in demand forecasts using adjusted exponential forecast

(E) Simple Linear Regression Method

This is a mathematical technique that relates to one variable, that is, the independent variable,
with another variable called the dependent in the form of a linear equation. The linear regression
equation is
y = a + b⋅ x
where y is the dependent variable, a is the intercept, b is the slope of the line and x is the dependent
variable. The variables a and b are given by

a = y − bx

b=
∑ xy − n( y )( x )
∑ x − n( x ) 2 2

where a is a constant, b is a coefficient of variable x, and x is the mean value of x, y is the mean
value of y, x is time, y is the demand, and n is the period for which data is analysed using linear
regression methods.
Coefficient of correlation shows the strength of correlation between two variables. The linear
regression model of forecasting is only used when the value of the coefficient of correlation (r) is
high, i.e. near to 1. Coefficient of correlation (r) in a linear regression equation is the measure of
the strength of the relationship between the independent (time) and dependent variable (demand).
It is given by
n

∑(x t − x )( yt − y )
r= t =1
n n

∑(x
t =1
t − x )2 ∑( y
t =1
t − y )2

The values of r vary between –1 and +1 with a value of +1, indicating a strong linear relationship
between the variables and –1 showing the strong reciprocal relationship. Coefficient of
 Forecasting  65

determination (r) is computed by squaring r. This is an indication of the percentage of variation

in the dependent variable as a result of the behaviour of the independent variable.

Example 3.5: The weekly demands of a motorcycle by a retailer are shown in Table 3.7. Find
an equation of the regression line and estimate the demand for the 14th week.

Table 3-7: Weekly demand of the car

Week (X): 1 2 3 4 5 6 7 8 9 10 11 12
Demand (Y): 420 450 460 420 500 550 480 520 610 570 600 590

Solution:
We use the formulas
a = y − bx

b=
∑ xy − n( y )( x )
∑ x − n( x )
2 2

For the calculations, see Table 3.8.

Table 3-8: Actual demands and forecasts of the motorcycle

Week (X) Demand (Y) xy x2 Forecast
1 420 420 1 419.4
2 450 900 4 436.63
3 460 1380 9 453.86
4 420 1680 16 471.09
5 500 2500 25 488.32
6 550 3300 36 505.55
7 480 3360 49 522.78
8 520 4160 64 540.01
9 610 5490 81 557.24
10 570 5700 100 574.47
11 600 6600 121 591.17
12 590 7080 144 608.93
Σx = 78 Σy = 6170 Σxy = 42,570 Σx2 = 650

Substituting the values from Table 3.8, we get

b=
∑ xy − n( y )( x ) = 42570 − 12 × (6170 /12) × (78/12) = 17.23
∑ x − n( x )
2 2
650 − 12 × (78/12) 2

a = y − bx = 514.1667 − 17.23 × 6.5 = 402.17

66  Industrial Engineering and Management

Therefore, the regression equation is

y = 402.17 + 17.23 x
y14 = 402.17 + 17.23 × 14
= 643.39
700

650

600
Demand (Units)

550
Demand (Y)
500 Forecast

450

400
1 2 3 4 5 6 7 8 9 10 11 12 13
Weeks

Figure 3-7: Actual demands and forecasts of the motorcycle using linear regression

In Figure 3.7, we can observe the linear nature of variation in actual demand. But, the variation
is not exactly linear in nature. A straight line can be drawn covering the maximum points.
The above variation shows the strong correlation between weeks and demand. The relationship
will be exactly a straight line when the value of r becomes 1.

(F) Causal Method of Forecasting

Causal forecasting methods are based on a known or perceived relationship between the factor
to be forecasted and other external or internal factors. The following models are used as causal
forecasting methods:
1. Regression: A mathematical equation relates a dependent variable to one or more
independent variables that are believed to influence the dependent variable.
2. Econometric model: This model comprises a system of interdependent regression
equations that describe some sector of economic activity.
3. Input–output model: This model describes the flows from one sector of the economy to
another, and so predicts the inputs required to produce outputs in another sector.
4. Simulation modelling: Simulation modelling leads to mathematical and computer
simulation. The actual system or real world is simulated and a model is formulated
which is finally used to correlate the demand variation with the variation in other factors
affecting the demand such as income of consumer, price of the product, technological
change in product design, etc.
 Forecasting  67

3.3.1 Seasonal Adjustment

A seasonal pattern is a repetitive increase or decrease in demand. There are several methods
for reflecting seasonal patterns in time-series forecasts. A seasonal factor is used to adjust
for seasonal patterns. Seasonal factors are multiplied by the normal forecast to get seasonally
adjusted forecasts.
One method for determining seasonal factors is to divide the demand for each seasonal
pattern by the total annual demand:

Si = ∑D / ∑∑Di ij

where i shows the specific season or quarter of a year; j shows the number of years Si is seasonal
adjustment factor; ΣDi shows the summation of demand in ith quarter of the past years; and
ΣΣDij shows the total demand in the past years. This results in a seasonal factor between 0.0
and 1.0 that can be applied to any time-series method. This equation reflects that a factor of
total annual demand is assigned to each season. The seasonal factors are then multiplied by the
annual forecast of the demand to yield a forecast for each season/quarter.

Example 3.6: The quarterly demands of a raincoat from a shop for the years 2012, 2013 and
2014 are given in Table 3.9. Forecast the demand for the year 2015 quarterly.

Table 3-9: Quarterly demands of the machine

Year Demand Total
Quarter 1 Quarter 2 Quarter 3 Quarter 4
2012 176 136 113 225 650
2013 191 153 125 232 701
2014 203 156 131 246 736
Total 570 445 369 703 2087

Solution:
We have

S1 = ∑D / ∑∑D
1 ij = 570 / 2087 = 0.273
S2 = ∑D / ∑∑D
2 ij = 445 / 2087 = 0.213
S3 = ∑D / ∑∑D
3 ij = 369 / 2087 = 0.176
S4 = ∑D / ∑∑D
4 ij = 703 / 2087 = 0.336

Multiply the forecasted demand for entire year by seasonal factors to determine quarterly
demand.
68  Industrial Engineering and Management

Forecast for the entire year (trend line for data in Table 3.9):

b=
∑ xy − n( y )( x ) = 4260 − 3 × (695.66) × (2) = 43.02
∑ x − n( x )
2 2
14 − 3 × ( 2)
2

So,
a = y − bx = 695.66 − 43.02 × 2 = 609.62

The regression equation is

y = 609.62 + 43.02 x

= 609.62 + 43.02 × 4
= 781.7
Seasonally adjusted forecasts:

SF1 = ( S1 ) ( F5 ) = (0.273)(781.7) = 213.4

SF2 = ( S2 ) ( F5 ) = (0.213) (781.7) = 166.5
SF3 = ( S3 )( F5 ) = (0.176 )(781.7) = 137.5
SF4 = ( S4 ) ( F5 ) = (0.336 )(781.7) = 262.6

Multiplicative Seasonal Method

To deal with seasonal effects in forecasting, two major parts of forecasting are to be completed,
as described below:
1. A forecast for the entire period (i.e. year) must be made using whatever forecasting
technique is appropriate.
2. This forecast will be developed to reflect the seasonal effects in each period (i.e. month
or quarter).
The multiplicative seasonal method adjusts a given forecast by multiplying the forecast by a
seasonal factor. The following steps are used to calculate the seasonal adjusted forecasting:
Step 1: Calculate the average demand Davg, t per period for each year (t) from the past data by
dividing the total demand for the year (ΣDp) by the number of periods (p) in that year (t).
Step 2: Divide the actual demand DP, t for each period (p) by the average demand Davg, t per
period (calculated in Step 1) to get a seasonal factor fp, t for each period; repeat for each
year of data.
Step 3: Calculate the average seasonal factor fp for each period by summing all the seasonal
factors Σfp, t for that period and dividing by the number of seasonal factors.
Step 4: Determine the forecast for a given period in a future year by multiplying the average
seasonal factor fp by the forecasted demand in that future year.
 Forecasting  69

Example 3.7: Solve Example 3.6 using the multiplicative method of seasonal adjustment.

Solution:

Table 3-10: Actual quarterly demand of a machine for the years 2012, 2013 and 2014
Year Demand Total Avg
Quarter 1 Quarter 2 Quarter 3 Quarter 4
2012 176 136 113 225 650 162.5
2013 191 153 125 232 701 175.25
2014 203 156 131 246 736 184.0

Table 3-11: Seasonal factor for forecasting of the machine for the years 2012, 2013 and 2014
Year Demand
Quarter 1 Quarter 2 Quarter 3 Quarter 4
2012 176/162.5 = 1.08 136/162.5 = 0.821 113/162.5 = 0.695 225/162.5 = 1.38
2013 191/175.25 = 1.09 153/175.25 = 0.873 125/175.25 = 0.713 232/175.25 = 1.32
2014 203/184 = 1.1 156/184 = 0.847 131/184 = 0.711 246/184 =1.33
Avg. seasonal 1.09 0.847 0.706 1.34
factor

Annual forecast for the year 2015 is predicted to be 782 units.

The average forecast per quarter is 782/4 = 195.5 units.
Now,
Quarterly forecast = Avg. forecast × Seasonal factor
Therefore,
Q1 : 1.09 (195.5 ) = 213.09
Q2 : 0.847 (195.5 ) = 165.58
Q3 : 0.706 (195.5 ) = 138.02
Q4 : 1.34(195.5) = 261.97

3.4 Forecasting Performance Measurement

Forecasting performance can be measured through the following terms:
Mean absolute deviation: Mean absolute deviation (MAD) is the average deviation of
forecasting of actual demand. It can be calculated as:
1 n
MAD = ∑ Dt − Ft
n t =1
70  Industrial Engineering and Management

where Dt is the actual demand for the period t and Ft is the forecast for the period t, and n is the
total number of periods.
Mean absolute percentage error (MAPE): It is very similar to MAD, but it is shown in
percentage. The MAPE can be calculated as:

100 n Dt − Ft
MAPE = ∑
n t =1 Dt
where Dt is the actual demand for the period t and Ft is the forecast for the period t, and n is the
total number of periods.
Mean square error: Mean square error (MSE) is used to show the small deviation at larger
scale by squaring the deviation. It can be calculated as:

1 n
MSE = ∑ ( Dt − Ft )2
n t =1
where Dt is the actual demand for the period t and Ft is the forecast for the period t, and n is the
total number of periods.

Tracking Signal
Tracking signal monitors any forecasts that have been made in comparison with actual, and warn
when there are unexpected departures of the outcomes from the forecasts. The tracking signal is
a simple indicator that forecast bias is present in the forecast model. It is most often used when
the validity of the forecasting model might be in doubt.
n

∑ (Actual demand − Forecast demand )i

Tracking signal = t =1

MAD
n

∑ (D t − Ft )
= t =1

MAD
where Dt is the actual demand for the period t and Ft is forecasting for the period t, and n is the
total number of periods.

Example 3.8: Using Table 3.12, find MAD, MAPE and MSE.

Table 3-12: Forecast errors

Week (X) Demand (D) Forecast (F) |Dt – Ft| |(Dt – Ft) /Dt| × 100 (Dt – Ft )2
1 420 419.4 0.6 0.142 0.36
2 450 436.63 13.37 2.971 178.75
3 460 453.86 6.14 1.334 37.69
4 420 471.09 51.09 12.164 2610.18
(Continued)
 Forecasting  71

Table 3-12: (Continued)

Week (X) Demand (D) Forecast (F) |Dt – Ft| |(Dt – Ft) /Dt| × 100 (Dt – Ft )2
5 500 488.32 11.68 2.336 136.42
6 550 505.55 44.45 8.081 1975.80
7 480 522.78 42.78 8.912 1830.12
8 520 540.01 20.01 3.848 400.40
9 610 557.24 52.76 8.649 2783.61
10 570 574.47 4.47 0.784 19.98
11 600 591.17 8.83 1.471 77.96
12 590 608.93 18.93 3.208 358.34
Σ = 275.11 Σ = 53.86 Σ = 10409.61

Solution:
1 n 275.11
MAD = ∑ Dt − Ft = 12 = 22.92
n t =1

100 n Dt − Ft 53.86
MAPE = ∑
n t =1 Dt
=
12
= 4.48 per cent

1 n 10409.61
MSE = ∑ ( Dt − Ft )2 = 12 = 867.46
n t =1

Summary
In this chapter, we have studied about forecasting, and its various models. Forecasting is an
important part of production planning which affects the various industrial or production activities
as discussed in purposes of forecasting for different time horizons. Forecasting is always wrong,
but its accuracy depends on the time horizon and analysis of the past data and present market
scenario. We have discussed about both the qualitative and quantitative forecasting. Qualitative
forecasting is normally used for the new product whose past data are not available and quantitative
forecasting involves the past data assuming that the past will be repeated in future.

Multiple-Choice Questions
1. In trend-adjusted exponential smoothing, the trend-adjusted forecast consists of
(a) the old forecast adjusted by a trend factor
(b) the old forecast and a smoothed trend factor
(c) an exponentially smoothed forecast and a smoothed trend factor
(d) an exponentially smoothed forecast and an estimated trend value
72  Industrial Engineering and Management

2. The primary method for associative forecasting is

(a) Delphi method (b) executive consensus
(c) regression analysis (d) exponential smoothing
3. The mean absolute deviation (MAD) is used for
(a) estimating the trend line (b) eliminating the forecast errors
(c) measuring the forecast accuracy (d) Seasonal adjustment
4. Customer service levels can be improved by
(a) sampling plan (b) control charting
(c) short-term forecast accuracy (d) customer selection
5. In business, forecasts are the basis for
(a) production planning (b) budgeting
(c) sales planning (d) all of the above
6. The two basic classifications of forecasting are
(a) mathematical and statistical (b) judgemental and qualitative
(c) historical and associative (d) qualitative and quantitative
7. Which of the following is not a type of judgemental forecasting?
(a) executive opinions (b) sales force opinions
(c) the Delphi method (d) time series analysis
8. A series of questionnaire is used in
(a) expert opinion (b) sales force opinion
(c) time series analysis (d) the Delphi method
9. In which of the following forecasting techniques, the last period actual demand is used as the
forecasting for the current period?
(a) Naïve method (b) Moving average method
(c) Exponential smoothing method (d) Regression method
10. Gradual and long-term movement in time series data is called
(a) cycles (b) seasonal variation
(c) trend (d) None of these
11. The basic difference between seasonality and cycles is
(a) the duration of the repeating patterns in seasonality is longer than cycles
(b) the duration of the repeating patterns in seasonality is shorter than cycles the magnitude of the
variation
(c) the duration of the repeating patterns in seasonality may be longer or shorter than the cycles
depends on the product
(d) none of these
12. Smoothing constant in the Naïve method is
(a) 0.2 to 0.5 (b) 0.5 to 0.7
(c) 1 (d) None of these
 Forecasting  73

13. Moving average forecasting techniques is used for

(a) immediately reflect changing patterns in the data
(b) lead changes in the data
(c) smoothening the variations in the data
(d) operate independently of recent data.
14. Which of the following forecasting techniques is based on the previous forecast plus a fixed fraction
of the forecast error?
(a) naive forecast
(b) simple moving average forecast
(c) exponentially smoothed forecast
(d) regression method
15. In an exponential smoothing method,
(a) all the past data have equal weights
(b) recent data has lesser weight than the past data
(c) recent data has more weight than the past data
(d) none of these

Answers
1. (c) 2. (c) 3. (c) 4. (c) 5. (d) 6. (d) 7. (d) 8. (d) 9. (a)
10. (c) 11. (b) 12. (c) 13. (c) 14. (c) 15. (c)

Review Questions
1. What is the objective of forecasting? Discuss the various steps involved in demand forecasting.
2. Explain the Delphi method of forecasting. How is it different from executive opinion method?
3. How does the weighted moving average method overcome the limitations of the moving average
method?
4. How does the weighted average method differ from the exponential smoothing method regarding the
weight assignment to the recent data?
5. What do you mean by seasonal variations in demand? How do you account the seasonality in
forecasting problems?
6. What are the techniques used to find the forecasting errors? Explain in detail.
7. What are the advantages of regression method of forecasting over simple moving average method of
forecasting?

Exercises
1. A company finds the relationships between the demand and economic index for the past 10 years
as shown in Table 3.13. (a) Determine the coefficient of correlation between these two variables and
(b) Determine the equation of the line of best fit. (c) Find the demand corresponding to economic
index 130.
74  Industrial Engineering and Management

Table 3-13: Demands and corresponding economics Indices for last 10 years
S. no. Demand Economic index S. no. Demand Economic index
1. 420 106 6. 540 111
2. 380 103 7. 720 122
3. 460 108 8. 280 100
4. 300 101 9. 180 92
5. 240 97 10. 400 105

2. The quarterly demands of three years 2012, 2013 and 2014 are given below (Table 3.14). Calculate
the quarterly demands for the year 2015.

Table 3-14: Quarterly demand

Year Quarter I Quarter II Quarter III Quarter IV
2012 250 400 470 340
2013 310 400 550 390
2014 420 520 600 480

3. The monthly demands for office furniture (in units) are given in Table 3.15. Forecast the demand
using 3-period and 5-period simple moving average for the 12th month. Also, show the variation in
forecasting graphically.

Table 3-15: Monthly demands for furniture for 11 months

Month (t) 1 2 3 4 5 6 7 8 9 10 11 12

Demand (Dt) 670 698 740 805 879 940 870 778 912 940 809 —

4. Using the data shown in Table 3.15, forecast the demand for the periods with the three-period weighted
moving average method. The most recent data should be given 50 per cent weightage, second year
past data 30 per cent, and third year past data 20 per cent. Also, compare the result with the result
of three-period simple moving average graphically.
5. Using the data shown in Table 3.15, forecast the demand for the periods using the exponential
smoothing method (α = 0.3 and α = 0.5). Also, compare the results graphically.
6. Using the data shown in Table 3.15, forecast the demand for the periods with adjusted exponential
smoothing method (α = 0.5 and β = 0.3). Also, compare the result graphically with simple exponential
smoothing (α = 0.5).
7. The weekly demands of a motorcycle by a retailer are shown in Table 3.16. Find an equation of the
regression line and estimate the demand for the 14th week.
Table 3-16: Weekly demand of the motor cycle
Week (X) 1 2 3 4 5 6 7 8 9 10 11 12

Demand (Y) 500 530 540 500 580 630 560 600 690 650 680 670
 Forecasting  75

 References and Further Readings

1. J. S. Armstrong, Fred Collopy, and J. Thomas Yokum (2005), ‘Decomposition by Causal Forces:
A Procedure for Forecasting Complex Time Series,’ International Journal of Forecasting, 21, 25–36.
2. James Golden, John Milewicz, and Paul Herbig (1994), ‘Forecasting: Trials and Tribulations’, Management
Decision, 32(1): 33–36.
3. Korpela, J. and Tuominen, M. (1996), ‘Inventory Forecasting with a Multiple Criteria Decision Tool’,
International Journal of Production Economics, 45, 159–168.
4. Kumar, P. (2012), Fundamentals of Engineering Economics (Delhi, Wiley India Pvt. Ltd).
5. Lancaster, G. A. and R. A. Lomas (1985), Forecasting for Sales and Material Management (London:
Macmillan).
6. Mentzer, J. T. and Moon, M. (2005), Sales Forecasting Management: A Demand Management Approach
(Thousand Oaks, CA; Sage Publications).
.
7. Pilinkiene , V. (2008), ‘Selection of Market Demand Forecast Methods: Criteria and Application’,
Inzinerine Ekonomika-Engineering Economics, 3(58): 19–25.
8. Waddell, D. (1994), ‘Forecasting: The Key to Managerial Decision-Making’, Management Decision,
32(1): 41–49.
9. Wisniewski, M. (2006), Quantitative Methods for Decision Makers (4th edition) (Harlow, Prentice
Hall).
 Chapter         4
Aggregate Planning

4.1 Introduction
Aggregate production planning is a planning in which general levels of employment and output
are planned to balance supply and demand typically for periods of one year or less. The term
‘aggregate’ implies for groups of products, or product types (i.e., product ‘families’) rather than
for specific or individual products. It is intermediate-range capacity planning, used to establish
employment levels, output rates, inventory levels, subcontracting and backorders for products
that are aggregated, i.e. grouped or brought together. It is not specifically focused on individual
products, but deals with the products in the aggregate.
Aggregate planning is essentially a broader approach to planning. Planners generally try to
avoid focusing on individual products or services unless, of course, the organization has only
one major product or service. Instead, they focus on overall, or aggregate, capacity. Aggregate
planning is closely related to other corporate decisions involving, for example, budgeting,
personnel and marketing. Most budgets are based on assumptions about aggregate output,
personnel levels, inventory levels, purchasing levels, etc. An aggregate plan should thus be the
basis for initial budget development and for budget revisions as conditions warrant (Pan and
Kleiner 1995).
Aggregate planning identifies the best way to utilize the limited resources of a firm to meet
the fluctuating demand. It simultaneously establishes optimal production, inventory and release
levels over a given finite planning horizon to meet the total demand (Buffa and Taubert 1972;
Hax and Candea 1984). The first production planning research was conducted by Hax and Meal
(1975) and aimed to find feasible solutions to planning. Axsater (1985) divided the planning
into two parts: aggregate planning and detailed planning. Unlike detailed planning, aggregate
planning is performed at higher level. Aggregate planning does not create restrictions for detailed
planning while at the same time considers long-term constraints.
Objectives of aggregate planning are to
(a) minimize costs of production/maximize profits
(b) maximize customer service
(c) minimize inventory investments
(d) minimize changes in production rates
(e) minimize changes in workforce levels
(f) maximize utilization of plant and equipment
 Aggregate Planning  77

4.2 Aggregate Planning Strategies

The aggregate planning problem can be clarified by a discussion of the various decision options
available. These will be divided into two types:
1. The options, influencing demand; and
2. The options, influencing supply.
Demand can be modified or influenced in several ways as mentioned below:
• Pricing: Price reduction leads to an increase in the demand and increase in price leads to
a decrease in the demand subject to type of goods.
• Advertising and promotion: Advertising may increase the demand by offering discounts,
additional offer, free post-sales service, etc.
• Back orders: It can be achieved by short selling, sell now and deliver later.
• New demand: New demand can be created by alternative use of the product.
There are also a large number of variables available to modify supply through aggregate planning
(Stevenson 1986). These include: hiring and layoff of employees, using overtime and undertime,
using part-time or temporary labour, carrying inventory, subcontracting, and making cooperative
arrangements.
On the basis of the above discussion, we observe that there are two strategies to meet the
fluctuating demand in the market. These strategies are: level strategies and chase strategies. In the
case of level strategy, production is uniform whereas in case of chase strategy, production chases
the demand by fluctuating the workforce utilization. Organizations must compare workforce
fluctuation costs with inventory costs to decide which strategy to use. Level strategy is used
when inventory costs are low as compared to the costs of fluctuating the workforce and when
efficient production is the primary goal. When inventory costs are high as compared to workforce
fluctuation costs, chase strategy is used, although it is less efficient for production. Apart from the
production strategies such as level and chase strategies, we also have planning strategies. There
are three major strategies associated with aggregate planning (Chase and Aquilano 1985):
• Production variations due to hiring, firing, overtime or undertime,
• Permitting inventory levels to vary, and
• Subcontracting
The demand of a product may continuously fluctuate with the market. The management has to
respond to the changes in demand by absorbing the demand fluctuations by carrying or allowing
inventory, adjusting the workforce through hiring and firing, and adjusting the production rate
through overtime and undertime, combining these alternatives. Each of these alternatives is
associated with some costs. The main purpose of aggregate planning is to minimize the all costs
over the planning horizon.
For example, forecasts of the expected demand of a product for the next six months and the
production days available during these months of planning horizon are given in Table 4.1. Also,
the costs associated with various production factors are shown in Table 4.2. Now, we want to find
the total production costs if we select different strategies of aggregate planning.
78  Industrial Engineering and Management

Table 4-1: Expected demands and the production days for the next six months

Month Expected demands Production days Demand per day

(computed)
Jan 1200 22 54
Feb 1000 15 67
Mar 1100 21 53
Apr 1400 25 56
May 1200 22 54
June 1600 20 80
Total 7500 125

Table 4-2: Costs associated with various production factors

Cost information
Inventory carrying cost Rs 4 per unit per month
Subcontracting cost per unit Rs 8 per unit
Average pay rate Rs 5 per hour (Rs 40 per day)
Overtime pay rate Rs 8 per hour (above 8 hours per day)
Labour-hours to produce a unit 1.8 hours per unit
Cost of increasing daily production rate Rs 320 per unit
(hiring and training)
Cost of decreasing daily production rate (layoffs) Rs 540 per unit

A. Constant Workforce Level and Carrying the Inventory

Total expected demand during planning horizon
Average total production =
Toal number of working days available during planning horizon
7500
= = 60 units per day
125
In this case, we maintain the constant production level equal to average production required per
day. In Figure 4.1, it can be observed that the constant production level is 60 units per day, while
the expected demand is continuously fluctuating from one month to another. Finally, at the end
of the planning horizon, the cumulative inventory becomes zero.
Table 4.3 shows the status of changes in inventory in each month and ending inventory,
i.e. cumulative inventory at the end of each month. We observe that at the end of the sixth
month, the total inventory becomes zero that means the demand in peak period is adjusted by the
inventory of the slack period. But, the total inventory carried over one month to another is equal
to 1000 units and in this case, total production costs will be equal to the normal production cost
plus inventory carrying cost of 1000 units in a month.
 Aggregate Planning  79

90
80
70
Demands 60
50
40
30
20
10
0
Working days 22 15 21 25 22 20
Months Jan. Feb. Mar. Apr. May June
Months and Working days

Figure 4-1: Constant production level equal to average expected demand per day

Table 4-3: Monthly change in inventory and cumulative or ending inventory

Month Production Production at Demand Monthly inventory Ending inventory

days 60 units per day forecast change (Cumulative)
Jan 22 1320 1200 +120 120
Feb 15 900 1000 −100 20
Mar 21 1260 1100 +160 180
Apr 25 1500 1400 +100 280
May 22 1320 1200 +120 400
June 20 1200 1600 −420 0
Total inventory for one month = 1000

Labour-hours required to produce 60 units per day = 60 × 1.8 = 108

108
Number of labours required per day = = 13.5
8
Labour cost per day = 13.5 × 5 × 8 = Rs 540
Total labour cost for 125 working days = 125 × 540 = Rs 67,500
Inventory carrying cost = 1000 × 4 = Rs 4000
Total production cost = Regular-time labour cost + Inventory carrying cost
= Rs 67,500 + Rs 4000 = Rs 71,500
Figure 4.2 shows the cumulative production and cumulative demand and how they become zero
at the end of the planning horizon.
80  Industrial Engineering and Management

8000

Cumulative production/demands
7000
6000
5000
4000 Cumulative
production
3000
Cumulative demand
2000
1000
0
Jan. Feb. Mar. Apr. May June
Months

Figure 4-2: Cumulative production and cumulative demand

B. Subcontracting as a Strategy to Meet the Fluctuating Demand

In the case of adopting subcontracting as a strategy to meet the fluctuating demand, the production
level remains constant equal to the minimum production level required, and any extra demand
occurs during the planning horizon is fulfilled by subcontracting. Figure 4.3 shows the constant
production level at minimum level equal to 53 units per day and the extra demands that are
fulfilled by subcontracting.
90
80
70
60
Demands

50
40
30
20
10
0
Working days 22 15 21 25 22 20
Months Jan. Feb. Mar. Apr. May June
Months and Working days

Figure 4-3: Meeting fluctuating demand by subcontracting

53 × 1.8
Regular-time labour cost for 125 working days = × 40 = Rs 59,625
8
Subcontracting units = 7500 − 53 × 125 = 875 units
 Aggregate Planning  81

Subcontracting cost = 875 × 8 = Rs 7000

Total production cost = Regular time labour cost + Subcontracting cost
= Rs 59,625 + Rs 7000 = Rs 66,625

C. Overtime as a Strategy to Meet the Fluctuating Demand

In the case of overtime production strategy, we maintain the production level at the minimum rate
required per day or the initial workforce available and the extra unit requirement is fulfilled by
overtime production. Here, we maintain the production rate of 53 units per day and if the demand
increases that amount is fulfilled by overtime production.
53 × 1.8
Regular-time labour cost for 125 working days = × 40 = Rs 59,625
8
Overtime units = 7500 − 53 × 125 = 875 units
Overtime cost = 875 × 1.8 × 8 = Rs 12,600
Total production cost = Regular time lab
bour cost + Overtime cost
= Rs 59,625 + Rs 12,600 = Rs 72,225

D. Hiring and Firing as a Strategy to Meet the Fluctuating Demand

In the case of hiring and firing, we have to produce the products as per the demand of the market.
Daily production rate fluctuates continuously. The cost of hiring and firing depends on the
increase and decrease in the daily production rate in the specific month, respectively. The total
cost of production will be equal to the regular time production cost plus hiring and firing costs of
the labour (Rs 89,580) as given in Table 4.4.
Table 4-4: Hiring and firing costs with regular-time labour cost

Month Forecast Daily Basic production Extra cost of Extra cost of Total cost
(units) production cost (demand × increasing decreasing
rate 1.8 hrs/unit × production production
Rs 5/hr) (hiring cost) (layoff cost)
Jan 1200 54 Rs 10,800 — — Rs 10,800
Feb 1000 67 Rs 9000 Rs 4160 — Rs 13,160
(= 13 × Rs 320)
Mar 1100 53 Rs 9900 — Rs 7560 Rs 17,460
(= 14 × Rs 540)
Apr 1400 56 Rs 12,600 Rs 960 — Rs 13,560
(= 3 × Rs 320)
May 1200 54 Rs 10,800 — Rs 1080 Rs 11,880
(= 2 × Rs 540)
June 1600 80 Rs 14,400 Rs 8320 — Rs 22,720
(= 26 × Rs 320)
Total Rs 67,500 Rs 13,440 Rs 8640 Rs 89,580
82  Industrial Engineering and Management

Table 4.4 shows the hiring and firing costs of labours with regular-time production cost,
and Figure 4.4 shows the aggregate planning with hiring and firing of labour. The total cost of
production using hiring and firing of labours as a strategy will be Rs 89,580.
90
80
70
60
Demands

50
40
30
20
10
0
Working days 22 15 21 25 22 20
Months Jan. Feb. Mar. Apr. May June
Months and Working days

Figure 4-4: Aggregate planning with hiring and firing

If we compare all the three strategies of aggregate planning, we find that with the given data,
subcontracting option with a total production cost of Rs 66,625 is the most suitable option.

4.3 Mixed Strategy

The model of mixed strategy is developed on the basis that all the four (discussed in previous
section) strategies are used optimally at the same time. For this model, we use linear programming
model and liner decision rule model.

4.3.1 Linear Programming for Aggregate Planning

If all the cost parameters in aggregate planning are linear, we can apply the linear programming
as a tool to solve the problem. Suppose following notations are given for the different cost
parameters:
cH = Cost of hiring one worker,
cF = Cost of hiring one worker,
ct = Cost of one unit in inventory for unit period,
cR = Cost off producing one unit on regular time,
cO = Incremental cost of producing one unit on overtime,
 Aggregate Planning  83

cU = Idle cost per unit of production,

cS = Cost of subcontract one unit of production,
nt = Number of production days in period t ,
K = Number of aggregate units produced by one worker in one day,
I O = Initial inventory on hand at the start of the planning horizon,
WO = Initial workforce at the start of the planning horizon,
Dt = Forecast of demand in time t .

The cost parameters may be time dependent. They may be useful for modelling changes in the
costs of hiring or firing due to shortages in the labour pool, or changes in the costs of production
or shortage due to poor or disrupted supply of resources, or change in interest rates. There are
some useful problem variables which are given below as:
Wt = Workforce level in period t ,
Pt = Production level in period t ,
I t = Inventory level in period t ,
H t = Number of workers hired in period t ,
Ft = Number of workers fired in period t ,
Ot = Overtime production in units,
U t = Worker idle time in units,
St = Number of units subcontracted from outside.
Suppose the total time of production is T; now the total costs involved in production during time
T will be:
cH H t (i.e., Hiring cost ) + cF Ft (i.e., Firing cost ) + cI I t (i.e., Inventory carrying cost )
+ cR Pt (i.e., Regular time production cost ) + co Ot (i.e., Overtime production cost )
+ cU U t (i.e., Idle production cost ) + cs St (i.e., subcontracting production cost)

The objective of linear programming is to minimize the total production costs, i.e. the objective
function is:
T
Minimize Z = ∑ (cH H t + cF Ft + cI I t + cR Pt + co Ot + cU U t + cs St )
t =1

Subject to constraints:
Wt = Wt −1 + H t − Ft ; 1≤ t ≤ T; Workforce constraints
I t = I t −1 + Pt + St − Dt ; 1≤ t ≤ T; Inventory constraints
Pt = KntWt + Ot − U t ; 1≤ t ≤ T; Productionn level constraints
84  Industrial Engineering and Management

Case I: No hiring and firing, no subcontracting, no overtime production, only constant workforce
and inventory are allowed
Total cost = cI I t (i.e., Inventory carrying cost ) + cR Pt (i.e., Regular time production cost )
+ cU U t (i.e., Idle production cost)
T
Thus, Minimize Z = ∑ (cI I t + cR Pt + cU U t )
t =1

Subject to the constraints:

Wt = Wt −1 ; 1≤ t ≤ T; Workforce constraints
I t = I t −1 + Pt − Dt ; 1≤ t ≤ T; Inveentory constraints
Pt = KntWt − U t ; 1≤ t ≤ T; Production level constraints

Case II: Only constant workforce and overtime production are allowed.

Total cost = cI I t (i.e., Inventory carrying cost ) + cR Pt (i.e., Regular time production cost )
+ co Ot (i.e., Overtime production cost ) + cU U t (i.e., Idle production cost )

The objective of linear programming is to minimize the total production costs, i.e. the objective
function is:
T
Minimize Z = ∑ (cI I t + cR Pt + co Ot + cU U t )
t =1

Subject to the constraints

Wt = Wt −1 ; 1≤ t ≤ T; Workforce constraints
I t = I t −1 + Pt − Dt ; 1≤ t ≤ T; Inveentory constraints
Pt = KntWt + Ot − U t ; 1≤ t ≤ T; Production level constraints

Case III: Only constant workforce and subcontracting are allowed.

Total cost = cI I t (i.e., Inventory carrying cost ) + cR Pt (i.e., Regular time production cost )
+ cU U t (i.e., Idle production cost ) + cs St (i.e., subcontracting production cost )
The objective of linear programming is to minimize the total production costs, i.e. the objective
function is:
T
Minimize Z = ∑ (cR Pt + cU U t + cs St )
t =1
Subject to the constraints
Wt = Wt −1t ; 1≤ t ≤ T; Workforce constraints
I t = I t −1 + Pt + St − Dt ; 1≤ t ≤ T; Inventory constraints
Pt = KntWt − U t ; 1≤ t ≤ T; Production level constraints
 Aggregate Planning  85

Case IV: Only hiring and firing are allowed.

Total cost = cH H t (i.e., Hiring cost ) + cF Ft (i.e., Firing cost )
+ cR Pt (i.e., Regular time production cost ) + cU U t (i.e., Idle production cost)
The objective of linear programming is to minimize the total production costs, i.e., the objective
function is:
T
Minimize Z = ∑ (cH H t + cF Ft + cR Pt + cU U t )
t =1

Subject to constraints:
Wt = Wt −1 + H t − Ft ; 1≤ t ≤ T; Workforce constraints
I t = I t −1 + Pt − Dt ; 1≤ t ≤ T; Inventory constraints
Pt = KntWt − U t ; 1≤ t ≤ T; Production level constraints

4.3.2 The Linear Decision Rule

The second approach to solve the aggregate programming problem is linear decision rule (LDR).
This rule was suggested by Holt et al. (1956). The costs involved in this model are inventory
costs, cost of changing production level, number of workers, etc. These costs are represented by
quadratic functions. The total cost over the period T can be given as:
T
Minimize Z = ∑ ⎡c1Wt + c2 (Wt − Wt −1 ) + c3 ( Pt − KntWt ) + c4 Pt + ( I t − c6 ) ⎤
2 2 2

t =1
⎣ ⎦

Subject to the constraints:

I t = I t −1 + Pt − Dt ; for 1 ≤ t ≤ T
Cost

Cost

Number of Number of Number units Number units

workers fired workers hired back ordered held

Figure 4-5: LDR for hiring and firing and backordering

The value of the constants c1, c2, c3, …, c6 must be determined for the specific applications. The
cost functions as quadratic functions (as shown in Figure 4.5) have some advantages over the
linear programming approach. As quadratic functions are differentiable, the standard rules of
calculus for maxima and minima can be applied to find the optimal solution.
86  Industrial Engineering and Management

Summary
In this chapter, we have discussed the various strategies to meet the fluctuating demand in the
intermediate range of planning. Level strategies are used to level the inventory as per demand,
but in chase strategies, the fluctuating demand is chased by various strategies such as overtime
production, subcontracting, changing the workforce level, etc. Also, we have discussed the linear
programming and LDR as a tool to formulate the mixed strategy.

Multiple-Choice Questions
1. Aggregate planning is capacity planning for
(a) the long range (b) the intermediate range
(c) the short range (d) typically one to three months
2. A chase demand strategy is used when
(a) material cost is high (b) labour cost is high
(c) inventory costs are high (d) all of the above
3. Which of the following aggregate planning method is an optimal solution?
(a) linear programming (b) search decision rule
(c) trial and error with a spreadsheet (d) none of these
4. Which of the following statements regarding aggregate planning for services is not true?
(a) Services cannot be inventoried (b) Demand is difficult to predict
(c) Capacity is easy to predict (d) Labour is a big constraint
5. Which of the following statements is not true for level production strategy of aggregate planning?
(a) Level production strategy sets production rate constant
(b) The main costs of level production involve hiring and firing
(c) Level production strategy uses inventory to absorb variations in demand
(d) All of the above
6. Level strategy is used when
(a) inventory costs are low as compared to the costs of fluctuating the workforce
(b) efficient production is the primary goal
(c) costs of subcontracting are high
(d) all of the above
7. Overtime as a strategy to meet the fluctuating demand is used when
(a) overtime cost is less than the inventory cost
(b) subcontracting cost is less than the inventory cost
(c) inventory cost is less than the cost of hiring and firing of the workers
(d) none of these
 Aggregate Planning  87

8. Hiring and firing as a strategy to meet the fluctuating demand is used under
(a) a level production strategy (b) a chase demand strategy
(c) both (d) none of these
9. Linear programming is used for
(a) only hiring and firing strategy (b) only constant workforce strategy
(c) only overtime production strategy (d) all the above and mixed strategy
10. Aggregate planning
(a) is used for only single product (b) may be used for a group of product
(c) is based on long range forecasting (d) none of these
11. Aggregate planning is used for
(a) the best way to utilize the limited resources of a firm
(b) minimizing the overall production cost
(c) meeting the fluctuating demand in the market
(d) all of the above
12. Which of the following strategies is used to adjust the capacity to match demand?
(a) using part-time labourers (b) changing price
(c) backordering (d) none of these
13. Which of the following strategies is known for lower employee morale?
(a) hiring and firing strategy (b) constant workforce strategy
(c) overtime production strategy (d) subcontracting strategy
14. Which of the following strategies is used to manipulate the demand of the product?
(a) hiring and firing strategy (b) constant workforce strategy
(c) overtime production strategy (d) pricing strategy
15. Which of the following strategies is a demand option?
(a) changing workforce level (b) changing inventory level
(c) changing production level (d) changing price level

Answers
1. (b) 2. (c) 3. (a) 4. (c) 5. (b) 6. (d) 7. (a) 8. (b) 9. (d)
10. (b) 11. (d) 12. (a) 13. (a) 14. (d) 15. (d)

Review Questions
1. What do you mean by aggregate planning? How does it differ from long-term planning?
2. How does the chase strategy differ from level strategy?
3. What are the actions taken for chasing the fluctuating demand?
88  Industrial Engineering and Management

4. What are the actions taken for levelling the production to meet the fluctuating demand?
5. What is the method to influence the fluctuating demand in the markets?
6. What are the limitations of hiring and firing options of workers?
7. In which situations subcontracting option is considered as the most suitable action?

Exercises
1. Forecasts of the expected demand of a product for the next six months and the production days
available during these months of planning horizon are given in Table 4.5. Also, the costs associated
with various production factors are shown in Table 4.6. Find the production costs under the following
strategies.
(a) Constant workforce level and carrying the inventory
(b) Subcontracting as a strategy to meet the fluctuating demand
(c) Overtime as a strategy to meet the fluctuating demand
(d) Hiring and firing as a strategy to meet the fluctuating demand

Table 4-5: Product demand and available production days

Month Expected demands Production days

Jan 1000 20
Feb 800 13
Mar 900 19
Apr 1200 23
May 1000 20
June 1400 18
7300 113

Table 4-6: Various costs involved in production

Cost information
Inventory carrying cost Rs 5 per unit per month
Subcontracting cost per unit Rs 8 per unit
Average pay rate Rs 5 per hour (Rs 40 per day)
Overtime pay rate Rs 10 per hour (above 8 hours per day)
Labour-hours to produce a unit 2 hours per unit
Cost of increasing daily production rate Rs 300 per unit
(hiring and training)
Cost of decreasing daily production rate (layoffs) Rs 500 per unit
 Aggregate Planning  89

 References and Further Readings

1. Axsäter, S. (1985), ‘On the Feasibility of Aggregate Production Plans’, Operations Research, 34(5):
796–800.
2. Buffa, E. S. and Taubert, W. H. (1972), Production-Inventory Systems Planning and Control
(Homewood, Illinois. Richard D. Irwin, Inc.).
3. Chase, R. B. and Aquilano, N. J. (1985), Production and Operations Management (Homewood, IL:
Irwin).
4. Hax, A. C. and Candea, D. (1984), Production and Inventory Management, Prentice Hall.
5. Hax, A. C. and Meal, H. C. (1975), ‘Hierarchical Integration of Production Planning and Scheduling’,
in M. A. Geisler (ed.), Studies in Management Sciences, (American Elsevier, New York: North Holland),
vol. 1. pp. 53–69.
6. Holt, C. C., Modigliani, F. and Muth, J. F. (1956), ‘Derivation of Linear Decision Rule for Production
and Employment’, Management Science, 2, 159–177.
7. Pan, L. and Kleiner, B. H. (1995), ‘Aggregate Planning Today’, Work Study, 44(3): 4–7.
8. Ravindran, A. (2008), Aggregate Capacitated Production Planning in a Stochastic Demand
Environment, Proquest, Dissertation at Purdue University.
9. Stevenson, W. J. (1986), Production/Operations Management (Homewood, IL: Irwin).
 Chapter         5
Capacity Planning: MRP,
MRP II and ERP

5.1 Introduction
Capacity planning is the process of determining the production capacity needed by an organization
to meet changing demands for its products. In the context of capacity planning, the capacity is
the maximum amount of work that an organization is capable of doing in a given period of time.
There are three basic steps of capacity planning: determination of service level or production
level requirements, analysis of the current capacity of the organization and planning for the
capacity required in future.
The required production level of product and services can be determined by finding the
workloads and identifying service level for each workload. The current capacity of the systems
can be analysed by measuring the production level and comparing with the target or objectives,
measuring the overall resource usage, measuring resource usage by workload and identifying the
components of response time. The future plan consists of future processing requirements and
plan for future system configuration and specifications.
Capacity requirement planning: The capacity requirement planning (CRP) is a solution to
meet the target of manufacturing organizations with discrete production systems. It allows the
maintenance of resources, tools, products, production plans and their interdependencies. CRP
supports the decision-making process with maximum user-friendliness, interconnectivity,
portability and flexibility. CRP first assesses the schedule of production that has been planned by
the company, and then analyses the company’s actual production capacity.
In this chapter, we will study about the material requirement planning (MRP), manufacturing
resource planning (MRP II) and enterprise resource planning (ERP).

5.2 Materials Requirement Planning

MRP was first introduced and used by IBM in 1970 and essentially substitutes excessive
inventories for better information systems. The MRP system was initiated in the 1960s and
was spearheaded by a team of IBM innovators comprising Joe Orlicky, George Plossl and Ollie
Wright, who sought to create a structured methodology for planning and scheduling materials
for complex manufactured products. MRP schedules the production of jobs through the factory
and eliminates the excessive work-in-process inventory levels required to compensate for job-
scheduling problems that arise in decoupled cost or operation centres. MRP releases work orders
for parts based upon a master production schedule and the current number and location of parts
 Capacity Planning: MRP, MRP II and ERP  91

within the plant. MRP is a system which maximizes the efficiency in the timing of raw materials
orders through to the manufacture and assembly of the final product.
MRP is a computer-based inventory management system designed to assist the managers
in scheduling and placing orders for items of dependent demand. Dependent demand items are
raw materials, component parts and subassemblies. For example, in a plant that manufactures
motorcycle, dependent demand inventory items might include rims, tires, seats and bike chains,
gearbox, engine blocks, electrical items, handle assembly items, break, clutches, etc.
MRP has been expanded to include information feedback loops so that production personnel
could change and update the inputs into the system as needed. The next generation of MRP,
known as manufacturing resources planning or MRP II, also incorporated marketing, finance,
accounting, engineering and human resources aspects in the planning process. A related concept
that expands on MRP is ERP (Enterprise Resource Planning), which uses computer technology
to link the various functional areas across an entire business enterprise.
MRP begins with a schedule for finished goods that is converted into a schedule of
requirements for the subassemblies, the component parts and the raw materials needed to produce
the final product within the established schedule. MRP is designed to answer three questions:
‘what is needed? How much is needed? when is it needed?’ The functions of MRP are to:
• create schedules and identify the specific parts and materials required to produce an end
item,
• determine exact number of components needed, and
• determine the dates when orders for those materials should be released, based on lead
times.
The entire MRP can be divided into three parts: inputs to MRP, processes and outputs of MRP
as shown in Figure 5.1.
INPUTS PROCESS OUTPUTS

Changes
Master
schedule Order releases
Planned order
schedules
Lead
time
MRP computer
programs Exception
reports
Bill of
materials Planning
reports
Performance
Inventory control reports
status

Figure 5-1: MRP Structures

92  Industrial Engineering and Management

5.2.1 MRP Inputs

The main sources of MRP inputs are a master schedule, lead times, a bill of materials and an
inventory record with present status. The bill of materials is a listing of all the raw materials,
component parts, subassemblies and assemblies required to produce one unit of a specific
finished product. The bill of materials is arranged in a hierarchy as shown in Figure 5.3, so that
managers can see what materials are needed to complete each level of production. MRP uses the
bill of materials to determine the quantity of each component that is needed to produce a certain
number of finished products. From this quantity, the system subtracts the quantity of that item
already in inventory to determine order requirements.
The master schedule outlines the planned production activities of the plant. Based on the
forecasting and the orders received from the customers/markets, it states the quantity of each
product that will be manufactured and the time frame in which they will be manufactured. The
master schedule divides the planning horizon into a number of weeks. The schedule covers a
time frame long enough to produce the final product. This total production time consists of the
sum of the lead times of all the related fabrication and assembly operations of the components
and subcomponents. Master schedules are generated according to demand and without regard to
capacity.
The inventory records file provides an accounting of how much inventory is already on hand
or on order, and thus should be subtracted from the material requirements. The inventory records
file is used to track information on the status of each item by time period. This includes gross
requirements, scheduled receipts and the expected amount on hand. It includes other details for
each item as well, like the supplier, the lead time and the lot size.

5.2.2 MRP Processing and Outputs

The MRP system determines the net requirements for raw materials, component parts and
subassemblies based on the information gathered from the bill of materials, master schedule
and inventory records file. It determines the gross material requirements, then subtracts out the
inventory on hand and adds back in the safety stock in order to compute the net requirements.
The main outputs from MRP include three primary reports and three secondary reports. The
primary reports consist of planned order schedules, order releases and changes to planned orders,
which might include cancellations or revisions of the orders. The secondary reports generated by
MRP include performance control reports (which are used to track problems like missed delivery
dates and stock outs in order to evaluate system performance), planning reports (which can be
used in forecasting future inventory requirements); and exception reports (which call managers’
attention to major problems).
The MRP process involves the following steps:
• Determine the gross requirements of a finished product.
• Determine the net requirements and when orders will be released for fabrication or
subassembly.
• Net requirements = Total requirements – Available inventory.
• Net requirements = (Gross requirements + Allocations) – (On hand) − Scheduled receipts.
 Capacity Planning: MRP, MRP II and ERP  93

Example 5.1: A firm manager receives an order of 500 staplers. The order is to be delivered
at the end of 10th week from now. He had 200 staplers in the inventory at the time of the order
receipts. The product components and its structure are shown in Figures 5.2 and 5.3, respectively.
The lead time to manufacture the components is shown in Figure 5.4. Using MRP, determine the
schedules for the orders to be released for the manufacturing and assemblies of the components
so that the 500 staplers can be delivered on time.

Rubber pad

Arm

Spring steel
and tooth
Pin
Carriage

Spring

Magazine
cartridge

Base

Rubber pad Strike pad

Figure 5-2: Stapler with its components

Level-0
Stapler

Top assembly Middle assembly Base assembly Level-1

Arm Rubber Spring Carriage Helical Magazine Pin Base Strike Rubber Level-2
pad steel and spring cartridge pad pad
Tooth

Figure 5-3: Product structure (Bill of materials) of staplers

94  Industrial Engineering and Management

Rubber pad
1 week
Top
spring steel and tooth assembly
1 week 1 week

Arm
Carriage 1 week

1 week

Magazine catridge
Middle
1 week assembly
Stapler
2 weeks 1 week
Helical spring
2 weeks
Pin
1 week Rubber pad
1 week
Base
Strike pad assembly
1 week
1 week
Base
1 week

0 1 2 3 4 5 6 7 8 9 10
week

Figure 5-4: Lead times of the various components of the stapler

Solution:
The order release schedules and the amount of the components at various stages of the
manufacturing are shown in Tables 5.1 to 5.14.

Table 5-1: MRP schedules for the stapler

Item: Stapler Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 500
Scheduled receipts 0
Projected on-hand inventory 200
Net requirements 300
Planned order receipts 300
Planned order releases 300
 Capacity Planning: MRP, MRP II and ERP  95

Table 5-2: MRP schedules for the top assembly of the stapler
Item: Top assembly Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300

Table 5-3: MRP schedules for the rubber pad (top assembly) of the stapler
Item: Rubber pad Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300

Table 5-4: MRP schedules for the spring steel and tooth (top assembly) of the stapler
Item: Spring steel and tooth Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300
96  Industrial Engineering and Management

Table 5-5: MRP schedules for the arm (top assembly) of the stapler
Item: Arm Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300

Table 5-6: MRP schedules for the middle of the stapler

Item: Middle assembly Lead time: 2 weeks
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300

Table 5-7: MRP schedules for the carriage (middle assembly) of the stapler
Item: Carriage Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300
 Capacity Planning: MRP, MRP II and ERP  97

Table 5-8: MRP schedules for the magazine cartridge (middle assembly) of the stapler
Item: Magazine catridge Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300

Table 5-9: MRP schedules for the helical spring (middle assembly) of the stapler
Item: Helical spring Lead time: 2 weeks
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300

Table 5-10: MRP schedules for the pin (middle assembly) of the stapler
Item: Pin Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300
98  Industrial Engineering and Management

Table 5-11: MRP schedules for the base assembly of the stapler
Item: Base assembly Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300

Table 5-12: MRP schedules for the rubber pad (base assembly) of the stapler
Item: Rubber pad Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300

Table 5-13: MRP schedules for the strike pad (base assembly) of the stapler
Item: Strike pad Lead time: 1 week
Week
1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300
 Capacity Planning: MRP, MRP II and ERP  99

Table 5-14: MRP schedules for the base of the stapler

Item: Base Lead time: 1 week

Week

1 2 3 4 5 6 7 8 9 10
Gross requirements 300
Scheduled receipts 0
Projected on-hand inventory 0
Net requirements 300
Planned order receipts 300
Planned order releases 300

5.2.3 Benefits and Limitations of MRP

The following are benefits of MRP:
1. Increased customer service and satisfaction.
2. Improved utilization of facilities and personnel.
3. Better inventory planning and scheduling.
4. Faster response to market changes and shifts.
5. Reduced inventory levels without reducing customer services.
The following are limitations of MRP:
1. Delays in scheduled receipts.
2. Changes in planned order sizes because of capacity constraints.
3. Changes in gross requirements which dictate changes in lot sizes at sub-component
levels.
4. Unavailability of raw materials for one sub-component may negate the need for a fellow
subcomponent as both must be ready for the parent production.
5. Utilization of the same parts at different levels, indicating the need to restructure the bill
of materials.

5.3 MRP II
In the 1980s, MRP technology was expanded to create a new approach called manufacturing
resource planning, or MRP II. ‘The techniques developed in MRP to provide valid production
schedules proved so successful that organizations became aware that with valid schedules other
resources could be better planned and controlled,’ Gordon Minty noted in his book Production
Planning and Controlling. ‘The areas of marketing, finance, and personnel were affected by
the improvement in customer delivery commitments, cash flow projections, and personnel
management projections.’
100  Industrial Engineering and Management

Minty says that ‘MRP II does not replace MRP, and also it is not an improved version
of MRP. Rather, it represents an effort to expand the scope of production resource planning
and to involve other functional areas of the firm in the planning process,’ such as marketing,
finance, engineering, purchasing and human resources. MRP II differs from MRP in that all of
these functional areas have input into the master production schedule. From that point, MRP is
used to generate material requirements and helps production managers plan the capacity. MRP II
systems often include simulation capabilities, so managers can evaluate various options.
MRP II integrates many areas of the manufacturing enterprise into a single entity for planning
and control purposes, from board level to operative and from five-year plan to individual shop-
floor operation. It builds on closed-loop MRP by adopting the feedback principle, but extending
it to additional areas of the enterprise, primarily manufacturing-related.
MRP II is defined by the American Production and Inventory Control Society (APICS) as
a method for the effective planning of all the resources of a manufacturing company (Higgins,
LeRoy and Tierney 1996). It is a new generation of the MRP system, which is a set of techniques
that uses bills of material, inventory, data and a master production schedule to calculate the
requirements for materials in a manufacturing company.

5.3.1 Characteristics of MRP II

The characteristics of MRP II can be summarized as follows (Higgins, LeRoy and Tierney 1996):
1. The operation and financial system are the same.
2. It has simulation capabilities that enable predictions to be made beforehand.
3. It involves every facet of the business from planning to execution.
4. MRP II offers a systematic method for planning and procuring materials to support
production.
5. MRP II is not a proprietary software system and can thus take many forms.
6. It is almost impossible to visualize an MRP II system that does not use a computer, but
an MRP II system can be based on either purchased / licensed or in-house software.
7. Almost every MRP II system is modular in construction.
8. Characteristic basic modules in an MRP II system are master production scheduling
(MPS), item master data, bill of materials (BOM), production resources data, inventories
and orders, purchasing management, MRP, shop-floor control (SFC), capacity planning
or CRP, standard costing, cost reporting/management, distribution resource planning
(DRP).
9. Some ancillary systems in MRP II are business planning, lot traceability, contract
management, tool management, engineering change control, configuration management,
shop-floor data collection, sales analysis and forecasting, Finite Capacity Scheduling
(FCS), general ledger, accounts payable (purchase ledger), accounts receivable
(sales ledger), sales order management, Distribution Requirements Planning (DRP),
warehouse management, project management, technical records, estimating, Computer-
Aided Design/Computer-Aided Manufacturing (CAD/CAM), Computer-Aided Process
Planning (CAPP).
 Capacity Planning: MRP, MRP II and ERP  101

5.3.2 Advantages of Using MRP II

The following are advantages of MRP II:
1. Better control of inventories
2. Improved scheduling
3. Productive relationships with suppliers
4. Improved design control
5. Better quality and quality control
6. Reduced working capital for inventory
7. Improved cash flow through quicker deliveries
8. Accurate inventory records
9. Timely and valid cost and profitability information

5.4 Enterprise Resource Planning

ERP is a management tool used for planning and monitoring all of the resources of a manufacturing
company, including the functions of manufacturing, marketing, finance and engineering (Wight
1993). ERP represents the application of the latest IT to the MRP II system. This is recognized as
being an effective management system (Ormsby et al. 1990) that has an excellent planning and
scheduling capability offering significant gains in productivity, dramatic increases in customer
service, much higher inventory turns and greater reduction in material costs.
ERP systems as comprehensive package software solutions that seek to integrate the
complete range of business processes and functions in order to present a holistic view of the
business from a single information and IT architecture (Gable 1998). Thus, ERP is seen as ‘an
integrated, m
­ ulti-dimensional system for all functions, based on a business model for planning,
control, and global (resource) optimization of the entire supply chain, by using state if the art
Information Systems/Information Technology (IS/IT) that supplies value added services to all
internal and external parties’ (Slooten and Yap 1999). The main ERP Critical Success Factors
(CSFs) fall under one of four main categories, namely, commitment from top management
(Devenport 1998), Business Process Reengineering (BPR), the IT infrastructure (Summer 1999),
and deploying change management.
Top management commitment: Management must be a part of ERP implementation and it has
been clearly demonstrated that for IT projects to succeed, top management support is critical.
Business process re-engineering: The concept of BPR was first introduced by Hammer in 1990.
BPR has been defined as a fundamental redesign of business processes to achieve dramatic
improvements in critical areas such as cost, quality, service and speed (Hammer 1990). BPR
began as a technique to help the organizations fundamentally rethink how they do their work in
order to dramatically improve customer service, cut operational costs, and become world-class
competitors. It has become a popular management tool for dealing with rapid technological and
business change in today’s competitive environment (Hamid 2004). With the rise of e-commerce,
enterprise systems, customer relationship management and other technology-enabled new
business practices, businesses now face major changes in much shorter time periods. The
102  Industrial Engineering and Management

challenges of the new Internet economy may offer an opportunity to apply the lessons learned
from a decade of BPR efforts, which likewise sought ways to manage major change (Alan et al.
2003). Implementing an ERP system involves reengineering the existing business processes to
the best business process standard (Gibson et al. 1999).
IT Infrastructure: Adequate hardware and networking infrastructure are required for ERP
application. An ERP system relies on its operation on sophisticated information technology
infrastructure. In addition to this infrastructure, clearly, the software configuration has a critical
influence on the implementation process and outcome (Holland and Light 1999).
Change management: One of the main obstacles facing ERP implementation is resistance to
change. ‘About half of ERP projects fail to achieve results because managers underestimate the
efforts involved in managing change’ (Pawlowski et al. 1999). To successfully implement ERP,
the way organizations do business will need to change and the ways people do their jobs will
need to change too (Koch et al. 1999). Thus, change management is essential for preparing a
company for the introduction of an ERP system, and its successful implementation.

5.4.1 Features of ERP

ERP has the following features:
1. It provides multi-platform, multi-facility, multi-mode manufacturing, multi-currency,
multi-lingual facilities.
2. It supports strategic, tactical and operational planning and execution activities.
3. It covers all functional areas like manufacturing, selling and distribution, payables,
receivables, inventory, accounts, human resources, purchases, etc.
4. It performs core activities and increases customer service, thereby improving the
corporate image.
5. It bridges the information gap across organizations.
6. It provides complete integration of systems across the departments and companies
under the same management.
7. It allows the automatic introduction of the latest technologies like Electronic Fund
Transfer (EFT), Electronic Data Interchange (EDI), Internet, Video conferencing,
E-commerce, etc.
8. It eliminates the problems like material shortages, productivity enhancements, customer
service, cash management, inventory problems, quality problems, on time delivery, etc.
9. It provides intelligent business tools like decision support system, executive information
system, data mining, etc.

5.4.2 Benefits of ERP

The areawise benefits of ERP are shown below as:
Operational benefits: Cost reduction, cycle time reduction, productivity improvement, quality
improvement, customer service improvement, etc.
 Capacity Planning: MRP, MRP II and ERP  103

Managerial benefits: Better resource management, improved decision-making and planning, and
performance improvement.
Strategic benefits: Support for business growth, support for business alliance, building business
innovations, building cost leadership, generating product differentiation, building external
linkages, enabling e-governance, sustaining competitiveness, etc.
IT infrastructure benefits: Building business flexibility for current and future changes, IT cost
reduction, increased IT infrastructure capability, etc.
Organizational benefits: Changing work patterns, facilitating organizational learning,
empowerment, building a common vision, shifting work focus, increased employee morale and
satisfaction.

 Summary
In this chapter, we have learnt about the capacity planning. We have discussed about MRP,
manufacturing resource planning, ERP and BPR. The main focus of the chapter was to minimize
the cost of the product by minimizing the inventory cost, improving the quality and providing the
product on time to the customer. Finally, we introduced the management tool, ERP, which is used
to integrate all the resources available in the enterprise.

Multiple-Choice Questions
1. MRP means
(a) Maximum retail price
(b) Material requirement planning
(c) Manufacturing resource planning
(d) Money resource planning
2. MRP is used for
(a) Dependent demand
(b) Independent demand
(c) Reserve stock
(d) Spare parts
3. Which of the following is NOT an input required for MRP?
(a) Bill of material
(b) Master production schedule
(c) Inventory on hand
(d) Financial status of the company
4. The hierarchy of components requires to produce a final product is represented by
(a) Bill of material (b) Components directory
(c) Lead time (d) Master file
104  Industrial Engineering and Management

5. MRP process involves the following steps

(a) Net requirements = (Gross requirements + Allocations) − (On hand) − Scheduled receipts
(b) Net requirements = (Gross requirements + Allocations) + (On hand) − Scheduled receipts
(c) Net requirements = (Gross requirements + Allocations) − (On hand) + Scheduled receipts
(d) None of these
6. MRP II
(a) does not replace MRP
(b) is not an improved version of MRP
(c) represents an effort to expand the scope of production resource planning and to involve other
functional areas of the firm in the planning process
(d) all the above
7. ERP is a management tool used for planning
(a) manufacturing activities
(b) marketing activities
(c) financial activities
(d) all the business activities
8. BPR (Business Process Reengineering)
(a) is a tool used to redesign the business processes to achieve dramatic improvements in critical
areas such as cost, quality, service and speed
(b) is a tool of ERP
(c) both is correct
(d) none of these
9. The demand for a given item is known as dependent if
(a) The item has many children
(b) There is a deep bill of materials
(c) The finished products are mostly services
(d) The item belongs to a clearly identifiable parent
10. The master production schedule represents the
(a) Financial requirements for the production
(b) Starting time components manufacturing
(c) Finishing time of components manufacturing
(d) Starting and finishing time of different products
11. The MPS calls for 120 units of product A. There are 20 units of product A on hand. One unit of A
requires 3 units of B and 5 units of C. There are 50 units of B and 130 units of C on hand. The net
requirement of B is
(a) 200
(b) 250
(c) 300
(d) 350
 Capacity Planning: MRP, MRP II and ERP  105

12. The MPS calls for 120 units of product A. There are 20 units of product A on hand. One unit of A
requires 3 units of B and 5 units of C. There are 50 units of B and 130 units of C on hand. The net
requirement of C is
(a) 250 (b) 300
(c) 350 (d) 370
13. Material requirement planning specifies
(a) the quantity of materials required to produce the products
(b) quantities and timings of planned order to be released
(c) capacity requirement to produce the product
(d) all the above
14. ERP is
(a) severely limited by MRP computer systems
(b) not related to MRP
(c) an advanced MRP-II systems integrated with customers and suppliers
(d) not related to MRP-II
15. Dependent and independent demand of an item differ in that
(a) for any items, the demands of all the components are dependent demand
(b) the quantity of independent demand is to be forecasted
(c) the quantity of dependent demand is to be calculated
(d) all the above

Answers
1. (b) 2. (a) 3. (d) 4. (a) 5. (a) 6. (d) 7. (d) 8. (c) 9. (d)
10. (d) 11. (b) 12. (d) 13. (b) 14. (c) 15. (d)

Review Questions
1. What do you mean by capacity resource planning? Discuss its importance in production management.
2. What is MRP? Discuss the various inputs required for MRP. What are the outputs of MRP?
3. How does the MRP differ from MRP-II?
4. What is ERP? What are the advantages of ERP?
5. Define BPR and discuss its utilization in ERP.
6. What are the influencing factors for the success of ERP implementation?

Exercises
1. A firm manager receives an order of 200 units of a product A. The order is to be delivered at the
end of 7th week from now. He had 50 units of product A in the inventory at the time of the order
receipts. The product structure is shown in Figure 5.5. The lead times to manufacture the components
106  Industrial Engineering and Management

are shown in Figure 5.6. Using MRP, determine the schedules for the orders to be released for the
manufacturing and assemblies of the components so that the 200 units of the product A can be
delivered on time.

Level Product structure for A

0 A

1 B(1) C(2)

2 D(1) E(2) F(2) G(2) H(3)

Figure 5-5: Product structure of A

D
E
B
F
A
G
C
H

1 2 3 4 5 6 7
Week

Figure 5-6: Lead time for the components

2. End item X is composed of three subassemblies: A, B and C. A is assembled using 3 Ds and 4 Es;
B is made of 2 Fs and 2 Gs; and C is made of 3 Hs. On-hand inventories are 20 Bs, 40 Ds and
200 Es. Scheduled receipts are 10 as at the start of week 3, 30 as at the start of week 6, and 200 Cs
at the start of week 3. One hundred Xs will be shipped at the start of week 6, and another 100 at
the start of week 7. Lead times are two weeks for subassemblies and one week for components D,
E and F. Final assembly of X requires one week. Include an extra 10 per cent scrap allowance in
each planned order of D. The minimum order size for E is 200 units. Develop each of the following:
(a) A product structure tree.
(b) An assembly time chart.
(c) A master schedule for X.
(d) A material requirements plan for A, D and E using lot-for-lot ordering.

 References and Further Readings

1. Alan, R., Traci, A., and Gigi, G. (2003), ‘Breaking the Rules: Success and Failure in Groupware-
supported Business Process Reengineering’, Decision Support Systems, 36(1): 31–47.
2. Davenport, T. (July–August 1998), ‘Putting the Enterprise into the Enterprise System’, Harvard
Business Review, 76(4): 121–131.
 Capacity Planning: MRP, MRP II and ERP  107

3. Gable, G. (1998), ‘Large Package Software: A Neglected Technology?’ Journal of Global Information
Management, 6(3): 3–4.
4. Gibson, N., Holland, C. P. and Light, B. (1999), ‘Enterprise Resource Planning: A Business Approach
to Systems Development’, 32nd Hawaii International Conference on System Sciences, Maui, HI, IEEE
Computer Society Press, Los Alamitos, CA, pp. 1–9.
5. Hamid R. A. (2004), ‘An Examination of the Role of Organizational Enablers in Business Process
Reengineering and the Impact of Information Technology’, Information Resources Management
Journal, 17(4): 1–19.
6. Hammer, M. (1990), ‘Reengineering Work: Don’t Automate, Obliterate’, Harvard Business Review,
68(4): 104–112.
7. Hasin, M. Ahsan A., and Pandey, P. C. (May–June 1996), ‘MRP II: Should its Simplicity Remain
Unchanged?’ Industrial Management, 38(3): 19.
8. Higgins, P., LeRoy, P., and Tierney, L. (1996), ‘Manufacturing Planning and Control: Beyond MRP II’
(London: Chapman & Hall).
9. Holland, C. and Light, B. (1999), ‘A Critical Success Factors Model For ERP Implementation’, IEEE
Software, 16(3): 30–36.
10. Koch, C., Slater, D. and Baatz, E. (December 22, 1999), ‘ABCs of ElU’, CIO Magazine.
11. Minty, Gordon (1998), Production Planning and Controlling (Tinley Park, IL: Goodheart-Willcox).
12. Ormsby, J. G., Ormsby, S. Y. and Ruthstrom, C. R. (1990), ‘MRP II Implementation: A Case Study’,
Production and Inventory Management Journal, 4: 77–80.
13. Pawlowski, S., Boudreau, M. and Baskerville, R. (1999), ‘Constraints and Flexibility in Enterprise
Systems: A Dialectic of System and Job’, In Proceedings of AMCIS, Milwaukee, WI, USA, 13–15
August.
14. Slooten, K. and Yap, L. (1999), ‘Implementing ERP Information Systems Using SAP’, In Proceedings
of AMCIS, Milwaukee, WI, USA, 13–15 August.
15. Stevenson, William J. (2002), Production/Operations Management. Seventh edition. (Irwin, McGraw-
Hill).
16. Sumner, M. (1999), ‘Critical Success Factors in Enterprise Wide Information Management Systems
projects’, In Proceedings of AMCIS, Milwaukee, WI, USA, 13–15 August.
17. Wight, O. (1993), The Executive’s Guide to Successful MRPI I, revised edition (Vermont: Oliver
Wight Publications), p. 1.
18. ‘Why SMEs Should Embrace MRP/ERP.’ Manufacturers’ Monthly. 16 March 2005.
 Chapter         6
Inventory Control

6.1 Introduction
Inventory is a stock of items kept on hand to meet the fluctuating demand. A certain level of
inventory is maintained that will help meet the anticipated demand. If demand is not known
with certainty, safety stocks are kept on hand. Additional stocks are sometimes used to meet
seasonal or cyclical demand. Sometimes, large amounts of items are purchased and stored to take
advantage of discounts.
In an organization, normally, four types of inventory are used as discussed below:
1. Raw materials and parts: These include all raw materials, components and assemblies
used in the manufacture of a product.
2. Consumables and spares: These may include materials required for maintenance and
day-to-day operations.
3. Work-in-progress inventory: These are items under various stages of production not yet
converted as finished goods.
4. Finished products: These are finished goods not yet sold or put into use.

6.1.1 Objectives of Inventory Control

The following are the objectives of inventory control:
1. To keep the investment in inventory to the minimum.
2. To minimize idle time by avoiding stock-outs and shortages.
3. To avoid carrying cost.
4. To improve customer satisfaction level with lesser inventory.
5. To avoid obsolescence of inventory.

6.1.2 Functions of Inventory

The following are the functions of inventory:
1. To meet anticipated demand: The demand cannot be estimated with full accuracy. There
is always chance of increasing or decreasing the demand due to various factors. To meet
this uncertainty in demand variation or to meet the anticipated demand, inventory is
required.
 Inventory Control  109

2. To smooth production requirements: There may be a sudden breakdown of the machine

involved in production. In this situation, an inventory is required at each workstation to
continue the production without any disruption.
3. To decouple operations: In case of product mix or multi-product production, we try to
postpone the product differentiation towards the end of the production process, so that
the demand of a customized product can be fulfilled very easily. Initially, a generic
product is produced and at the end of production; these products are stored in generic
form and finally customized as per demand. The point of customization or differentiation
is known as decoupling point.
4. To protect against stock-outs: A safety or reserve stock is required to meet the shortages
or delay in the replenishment of inventory. In case of shortages, production is disrupted
or demand of the customer remains unfulfilled which leads to poor customer satisfaction
level. Thus, inventory is required to protect against stock-out.
5. To help hedge against price increases: The price of a product may increase in near
future. In this case, inventory is required to help hedge against price increases.
6. To take advantage of quantity discounts: Inventory provides economies of scale. If we
produce or purchase the items in bulk or batch, there may be provision of quantity
discount.

6.1.3 Understocking and Overstocking

Understocking of inventory results in cost incurred when an item is out of stock. It includes the
cost of lost production during the period of stock-out and the extra cost per unit which might have
to be paid for an emergency purchase. Also, sometimes the company may incur an intangible cost
like opportunity cost and poor reputation in the market due to understocking.
Overstocking results in the cost incurred due to obsolescence and pilferage, and the
opportunity cost in the form of capital invested in inventory that could be invested for some
other purposes.

6.2 Classifications of Inventory

The methods used to analyse the items in inventory are discussed in the following subsections.
Always Better Control Analysis
As is evident from the above heading, ‘ABC’ connotes ‘Always Better Control.’ The basis of
analysing the annual consumption cost of the inventory items is the principle ‘Vital Few – Trivial
Many,’ and the criterion used here is the money spent, not the quantity consumed. Figure 6.1
shows the value and corresponding units in inventory. A-class items are very important and
have value up to 70 per cent, but their inventory used is up to 10 per cent. B-class items are
of moderate importance and have a value up to 20 per cent and also their inventory used is up
to 20 per cent. C-class items are of lower importance and have a value up to 10 per cent, but
inventory up to 70 per cent.
110  Industrial Engineering and Management

100 Class C
Class B
90

Percentage of Rs value
80
70 Class A
60
50
40
30
20
10
0
10 20 30 40 50 60 70 80 90 100
Percentage of SKUs

Figure 6-1: ABC analysis

Vital, Essential and Desirable Analysis

The inventory items, when subjected to analysis based on their criticality, can be classified into
vital, essential and desirable (VED) items. This analysis is termed as VED analysis.
1. Vital: Items without which production or a system comes to a standstill. The non-
availability of these items cannot be tolerated.
2. Essential: Items whose non-availability can be tolerated for 2–3 days because similar or
alternative items are available.
3. Desirable: Items whose non-availability can be tolerated for a long period.
Although the proportion of vital, essential and desirable items varies from company to company
depending on the type and quantity of workload, on an average vital items are 10 per cent,
essential items are 40 per cent and desirable items are 50 per cent of the total inventory available.

Fast, Slow-Moving and Non-moving Analysis

Fast, slow-moving and non-moving (FSN) analysis is based on the rate of consumption of
inventory items. The items can be classified as fast-moving, slow-moving and non-moving items.
An understanding of the movement of items helps in keeping proper levels of their inventory by
deciding a rational policy or reordering. Here, rational policy is concerned with decisions about
reorder point and order size as per consumption of the items. This method is based on the fact
that some stock items have a much higher annual usage than others.

Scarce, Difficult, or Easy Analysis

The classification of inventory items is based on their availability that could be scarce, difficult
or easy (SDE). The importance and storage of an item are decided on the basis of difficulty or
ease of its availability.
 Inventory Control  111

1. S refers to scarce items, especially imported ones and those which are very much in
short supply.
2. D refers to difficult items which are procurable in the market, but not easily available. For
example, items which have to come from far off cities or there is not much competition
for them in the market or their good quality supplies are difficult to get or procure.
3. E refers to easy items. These are mostly local items which are easily available.

6.3 Inventory Costs

The following costs are involved in keeping inventory:
1. Inventory carrying costs: Costs of holding items in storage. Important features of these
costs are as follows:
(a) These costs vary with the level of inventory and sometimes with the length of time
the inventory is held.
(b) These include facility operating costs, record-keeping, interest, etc.
(c) These are assigned on per unit basis per time period, or as percentage of average
inventory value (usually estimated as 10 to 40 per cent).
2. Ordering costs: Costs of replenishing the stock of inventory. Important features of these
costs are as follows:
(a) These costs are expressed as rupees per order, independent of the order size.
(b) These costs vary with the number of orders made.
(c) These costs include purchase orders, shipping, handling, inspection, etc.
3. Shortage (stock-out) costs: Costs associated with insufficient inventory. Important
features of these costs are as follows:
(a) These costs result in permanent loss of sales and profits for items not on hand.
(b) Sometimes, penalties are involved; if the customer is internal, work delays could
result, i.e., if the consumer is ready to bear the small delay the supplier may be
imposed with penalty due to shortages to minimize the loss.
The relationship between the ordering cost, inventory carrying cost and total cost is shown in
Figure 6.2.

Total cost
Annual cost

Minimum cost

Holding cost

Carrying cost

Lot size (Q)

Figure 6-2: Optimization of inventory costs

112  Industrial Engineering and Management

6.4 Continuous and Periodic Inventory Review Systems

The inventory review system is concerned with the volume of items in stock and its replenishment.
There are two types of review systems: continuous review and periodic review. In a continuous
review system, an order is placed for the same constant amount when inventory decreases to
a specified level. In a periodic review system, an order is placed for a variable amount over a
specified period of time.
In the continuous review system, the following process is followed.
1. The inventory level is reviewed continuously.
2. Whenever inventory decreases to a predetermined level, the reorder point, an order
is placed for a fixed amount to replenish the stock. The fixed amount is termed as the
economic order quantity (EOQ), whose magnitude is set at a level that minimizes the
total inventory carrying, ordering and shortage costs.
Because of continual reviewing, the management is always aware of the status of the inventory
level and critical parts, but this system is relatively expensive to maintain.
In the periodic review system, the following process is followed:
1. Inventory on hand is counted at specific time intervals and an order placed that brings
inventory up to a specified level. Inventory is not monitored between counts; therefore,
the system is less costly to track and keep account of inventory. The system results in
less direct control of the management and thus generally higher levels of inventory to
guard against stock-outs.
2. A new order quantity each time an order is placed. It is used in smaller retail stores, drug
stores, grocery stores and offices.

6.5 Economic Order Quantity

The EOQ and formula were originally presented by Ford Whitman Harris (1913) in his paper
‘Factory, The Magazine of Management’. The EOQ, or the economic lot size, is the quantity
ordered when inventory decreases to the reorder point. The amount is determined using the EOQ
model. The purpose of the EOQ model is to determine the optimal order size that will minimize
the total inventory costs.

Assumptions of the EOQ Model

The following are the assumptions of the EOQ model:
1. The demand rate is constant and is known with certainty.
2. No constraints are placed on the size of each lot.
3. The only two relevant costs are the inventory holding cost and the fixed cost per lot for
ordering or set-up.
4. Decisions for one item can be made independently of decisions for other items.
5. The lead time is constant and is known with certainty.
 Inventory Control  113

In this section, we will discuss the following four EOQ inventory models:
1. Basic EOQ model
2. EOQ model without instantaneous receipt
3. EOQ model with shortages
4. EOQ model with quantity discount

6.5.1 Basic EOQ Model

The following are the assumptions of the basic EOQ model:
1. Demand is known with certainty and is relatively constant over time.
2. No shortages are allowed.
3. Lead time for the receipt of orders is constant.
4. The entire order quantity is received at once and instantaneously.
In Figure 6.3, Q is maximum inventory level. At the beginning of the cycle, the inventory level
is Q and at the end of the cycle it becomes zero. Therefore, the average inventory level is Q/2.
The reorder point depends on the consumption rate and lead time. At reorder point, the inventory
should be able to meet the need during the lead time of replenishment/supply, i.e. reorder point
is equal to lead time multiplied by consumption rate. Lead time is the time interval between the
point of order placement and order receipt. In the basic model of EOQ, the supply is instantaneous
demand is constant throughout the year.
The inventory carrying cost is usually expressed on a per unit basis of time, traditionally one
year. The inventory carrying cost is calculated by multiplying inventory cost per unit per year
to average inventory size. Ordering cost per year can be calculated by multiplying the ordering
cost per order to the number of orders per year. The derivation of expression for EOQ (Q*) can
be given below as:

Demand
Order rate
quantity, Q
Inventory level

Q/2
Reorder
point, R

O Lead Lead Time

time time
Order Order Order Order
placed receipt placed receipt

Figure 6-3: The basic EOQ model

114  Industrial Engineering and Management

Inventory carrying cost per unit per year = Cc

Q
Average inventory =
2
Cc Q
Annual inventory carrying cost =
2

Total annual ordering cost equals cost per order (Co) times number of orders per year. Now
number of orders per year with known and constant demand is given by D/Q, where D is the
annual demand and Q is the order size. Also, we have
D
Annual ordering cost = Co ×
Q
D Q
Therefore, total inventory cost , TIC = Co × + Cc ×
Q 2
Differentiating TIC w.r.t. Q, we get minimum TC as:
dTIC
=0
dQ
2Co D
Q* =
Cc

Annual
cost
Total cost
Slope = 0

Minimum Carrying cost = Cc Q

2
total cost

Ordering cost = Co D
Q

Optimal order, Order quantity, Q

Qopt

Figure 6-4: Optimization of total inventory costs in basic EOQ model

Inventory carrying cost increases as the order size increases since it is calculated per unit of
item per unit time. But, the ordering cost decreases due to increase in order size because of less
number of orders. The total variable cost is the summation of carrying costs and ordering cost
as shown in Figure 6.4. At an optimum order size, the slope of total cost becomes zero and total
variable costs become minimum.
 Inventory Control  115

Example 6.1: A manufacturer deals exclusively with the production of brake shoes. Currently,
the company is trying to decide on an inventory and reorder policy for the brake shoes. A brake
shoe costs the company Rs 75 each, and demand is about 5000 brake shoes per year distributed
fairly evenly throughout the year. Reordering costs are Rs 800 per order and carrying costs are
figured at 20 per cent of the cost of the item. The company is open 300 days a year (6 days a
week and closed two weeks in August). The lead time is 40 working days. Find the EOQ, the
reorder point, the number of orders per year and total variable costs.

Solution:
We have Ci = Rs 75, D = 5000 units, Co = Rs 800, Cc = 0.2 × Ci = 0.2 × 75 = Rs 15, L = 40 days.
Now the EOQ is
EOQ is
2Co D 2 × 800 × 5000
Q* = = = 730.29 = 731 unit of brake shoes
Cc 15
Now, L = 40 days and daily demand d = 5000/300 = 16.67; therefore, the reorder point is
r = (16.67)( 40 ) = 666.66 ≈ 667 units
The company should reorder 731 brake shoes when his inventory position reaches 667. Number
of reorder times per year is
(5000 / 731) = 6.83
or after
(300 / 6.83) = 43.92 ≈ 44 working days
So
Total costs = (Carying cost) + (Ordering cost)
TC = [Cc (Q / 2)] + [Co ( D /Q )]
= [0.2(75)(Q / 2)] + [800(5000 /Q )]
= 15Q + ( 40, 00, 000 /Q )
Therefore,
TC = 15(731) + ( 40, 00, 000 / 731) = Rs 16, 436.95.

6.5.2 EOQ Model without Instantaneous Receipt

The second major paper was presented by Taft (1918) who developed the production lot size
model. This model is an extension of the simple EOQ model incorporating the production
rate. Camp (1922) was the first to present a general formula to determine the production order
quantity, such that the total cost per unit for setting up plus interest on stores investment would
be a minimum. The following are the assumptions of the EOQ model with instantaneous receipt:
116  Industrial Engineering and Management

6.5.3 Demand Occurs at a Constant Rate of D Items Per Year

1. The production rate is P items per year (and P > D).
2. Set-up cost: Rs Co per run.
3. Carrying costs: Rs Cc per item in inventory per year.
4. Purchase cost per unit is constant (no quantity discount).
5. Set-up time (lead time) is constant.
6. Planned shortages are not permitted.
In this model, the inventory replenishment is not instantaneous. Generally, this type of inventory
model is used in production of an item. Inventory replenishment is a continual process; similarly,
the inventory depletion is also a continual process. Both the processes occur simultaneously, but
the rate of inventory replenishment (i.e. p) is greater than inventory depletion (i.e. d) as shown in
Figure 6.5. The fraction of total production stored in stock is (p-d)/p or (1 − d/p). The maximum
level of inventory stored in stock is Q(1 − d/p) and average inventory level is Q(1 − d/p)/2. If p
is the daily rate at which the order is received over time and d is the daily rate at which inventory
is demanded, then
Inventory Inventory
replenished depleted
⎛ d⎛
Q 1– Maximum
Inventory level

⎝ p⎝
inventory level

Q⎛ d⎛
Average
1– ⎝
2⎝ p
inventory level

0 Begin End Time

Order
order order
receipt
receipt receipt
period

Figure 6-5: Graphical representation of the EOQ model without instantaneous receipt

⎛ d⎞
Maximum inventory level = Q ⎜1 − ⎟
⎝ p⎠
Q⎛ d⎞
Average inventory level = 1− ⎟
2 ⎜⎝ p⎠
D
Total ordering cost = Co
Q
Q⎛ d⎞
Total carrying cost = Cc 1− ⎟
2 ⎜⎝ p⎠
D Q⎛ d⎞
Total Inventory cost, TIC = Co + Cc ⎜ 1 − ⎟
Q 2⎝ p⎠
 Inventory Control  117

Differentiating TIC w.r.t. Q, we get minimum TC as:

dTIC D C ⎛ d⎞
= 0 = − Co 2 + c ⎜ 1 − ⎟
dQ Q 2 ⎝ p⎠
2Co D
or, Q* =
⎛ d⎞
Cc ⎜ 1 − ⎟
⎝ p⎠
Therefore,
2Co D
Optimal order size Q∗ =
Cc (1 − d /p)

Example 6.2: A cement manufacturing company has been using production runs of 1,00,000
packets, 10 times per year to meet the demand of 10,00,000 packets annually. The set-up cost
is Rs 5000 per run and the carrying cost is estimated at 10 per cent of the manufacturing cost
of Rs 100 per packet. The production capacity of the machine is 5,00,000 packets per month.
The factory is open 365 days per year. Calculate the optimal production lot size; the number
of production runs per year; the total annual variable cost; and the difference between existing
total annual variable cost and optimal annual variable costs; the order time between production
runs; the maximum inventory; and machine utilization.

Solution:
We have D = 1,000,000; Co = Rs 5,000; Cc = Rs 10; d = 1,000,000 per year, p = 12 × 5,00,000
= 6,000,000 per year
The optimal production lot size is given by
Q* = 2 DCo /[(1 − d /p)Cc ]
= 2(1, 000, 000)(5, 000) /[(10)(1 − 1/ 6)]
= 34, 641 packets
The number of production runs per year is
D /Q* = 28.86 times per year
Also,
34641 ⎛ 10, 00, 000 ⎞
Optimal TC = 5000(10, 00, 000 / 34641) + 10 × × ⎜1 − ⎟ = Rs 288, 675.13
2 ⎝ 60, 00, 000 ⎠
1, 00, 000 ⎛ 10, 00, 000 ⎞
Current TC = 5000(10, 00, 000 /1, 00, 000) + 10 × ⎜1 − ⎟ = Rs 466, 666.66
2 ⎝ 60, 00, 000 ⎠
Therefore, the difference is
466, 666.66 − 288, 675.13 = Rs 177, 991.53
118  Industrial Engineering and Management

There are 28.86 cycles per year. Thus, each cycle lasts
(365/28.86) = 12.64 days
The time to produce 34,641 packets per run is
(34, 641/ 6, 000, 000)365 = 2.10 days
Thus, the machine is idle for
(12.64 − 2.10) = 10.54 days between each run

6.5.4 EOQ Model with Shortages

The assumptions of the EOQ model with shortages are as follows:
1. Demand occurs at a constant rate of D items/year.
2. Ordering cost: Rs Co per order.
3. Holding cost: Rs Ch per item in inventory per year.
4. Backorder cost: Rs Cb per item backordered per year.
5. Purchase cost per unit is constant (no quantity discount).
6. Set-up time (lead time) is constant.
7. Planned shortages are permitted (backordered demand units are withdrawn from a
replenishment order when it is delivered).
In this model, Q is the maximum size of inventory; S is the inventory required or to be consumed
during the shortages period t2; Q − S is the inventory available for the period t1. Here, t2 indicates
the shortage period. Total cycle time is t as shown in Figure 6.6. Inventory exists only for
t1 period and for rest t2 period, there is a shortage of items.

Total units over time period

Time period in days =
Demand in units per day
Inventory level

Q–S
Q

0
Time
S

t1 t2

Figure 6-6: Graphical representation of the EOQ model with shortages

 Inventory Control  119

{( Average inventory level over t1 ) × t1

+ ( Average inventory level over t 2 ) × t 2 }
Average inventory level =
t
(Q − S )
t1 + 0 ⋅ t 2
= 2
t
(Q − S ) (Q − S )
⋅ + 0 ⋅ t2
= 2 D
Q
D
(Q − S ) 2
=
2Q

{(Average shortage level over t1 ) × t1

+ ( Average shortage level over t 2 ) × t 2 }
Average shortage level =
t
(S )
0 ⋅ t1 + ⋅t
= 2 2
t
S S
0+ ⋅
= 2 D
Q
D
2
(S )
=
2Q

S2
Total shortage cost = Cs
2Q
(Q − S ) 2
Total carrying cost = Cc
2Q
D
Total ordering cost = Co
Q
Total inventory cost = Orderring cost + Carrying cost + Shortage cost
S2 (Q − S ) 2 D
= Cs + Cc + Co
2Q 2Q Q
120  Industrial Engineering and Management

Differentiating the total inventory cost w.r.t. Q, we get

2Co D ⎛ Cs + Cc ⎞
Optimal order quantity, Q* = ⎜ ⎟
Cc ⎝ C s ⎠
⎛ Cc ⎞
Shortage level, S * = Q * ⎜ ⎟
⎝ Cc + C s ⎠

Slope = 0 Total cost

Cost

Minimum Carrying cost

cost

Order cost
Shortage cost
0
Qopt Q

Figure 6-7: Optimization of the total inventory costs in the EOQ model with shortages

In Figure 6.7, there are three variable costs: carrying cost, ordering cost and shortages cost.
These costs vary with order size, Q. Ordering cost and shortages cost decrease with increase in
order size, but carrying cost increases with increase in order size. Total cost is the summation of
these three costs and becomes minimum when its slope equals to zero. At zero slope of total cost,
the inventory size is known as optimal or EOQ.

Example 6.3: A car retailer has a monthly demand of 12 cars. Each car costs Rs 8,50,000. There
is also a Rs 10,000 order cost (independent of the number of cars ordered). The retailer has an
annual carrying cost at the rate of 5 per cent on each car. It takes two weeks to obtain cars after
they are ordered. For each week, if one car is out of the market, the retailer loses Rs 4000 profit.
(a) Find the EOQ and the optimal order policy. (b) How many days after receiving an order does
the retailer run with inventory? (c) How long is the retailer without any inventory per cycle?

Solution:
We have
D = 12 × 12 = 144 cars; Co = Rs 10, 000;
Cc = 0.05(8, 50, 000) = Rs 42, 500;
Cb = 4000 × 52 = Rs 2, 08, 000
 Inventory Control  121

Now,

(a) Q* = 2 DCo /Cc × (Cc + Cb ) /Cb

= 2(144)(10, 000) / 42, 500 × ( 42, 500 + 2, 08, 000) / 2, 08, 000 = 9.03

S * = Q *[Cc / (Cc + Cb ) ]
= 9.03[42, 500 / ( 42, 500 + 2, 08, 000)]
= 1.53
Demand is 12 cars per month or 3 cars per week. Since lead time is 2 weeks, lead time
demand is 6. Since the optimal policy is to order 9 cars and one car is out of market each
week. The order should be placed when there are 5 cars remaining in inventory.

(b) Inventory exists for Cb / (Cc + Cb ) = 2, 08, 000 / ( 42, 500 + 2, 08, 000))
= 0.83 of the order cycle
Note (Q * − S *) /Q* = 0.83 also, before Q * and S * are rounded. An order cycle is

Q * / D = 0.0627 years = 22.88 days

Thus, the retailer has inventory (0.83)( 22.88) = 19 days after receiving an order.

(c) Service is out of stock for approximately 22.88 − 19 = 3.88 days.

6.5.5 EOQ Model with Quantity Discount

The EOQ model with quantity discount is applicable when a supplier offers a lower purchase cost
for an item ordered in large quantities. In this model, variable costs are annual carrying, ordering
and purchase costs. For the optimal order quantity, the annual carrying and ordering costs are not
necessarily equal.
The following are the assumptions of the EOQ model with discount:
1. Demand occurs at a constant rate of D items/year.
2. Ordering cost is Rs Co per order.
3. Holding or carrying cost is Rs Cc = Rs CiI per item in inventory per year. (Note that the
holding cost is based on the cost of the item, Ci.)
4. The purchase cost is Rs C1 per item if the quantity ordered is between 0 and x1, Rs C2 if
the ordered quantity is between x1 and x2, and so on.
5. Delivery time (lead time) is constant.
6. Planned shortages are not permitted.
Total annual cost = Carrying + Ordering + Purchase
= [(1/ 2)Q * Cc ] + [ DCo /Q*] + DC
122  Industrial Engineering and Management

Example 6.4: An electronic shop carries film for the camera. The film costs the shop Rs 80
per unit and is sold for Rs 85. The shop has the film shelf life of 18 months. The shop sales
are 21 films per week, and its annual inventory carrying cost rate is 20 per cent. It costs the
shop Rs 20 to place an order. The film manufacturing company offers a 7 per cent discount on
orders of 400 films or more, a 10 per cent discount for 900 films or more and a 15 per cent
discount for 2000 films or more. Determine the shop’s optimal quantity.

Solution:
We have D = 21(52) = 1092; Cc = 0.20 (Ci); Co = 20.
(a) For C4 = 0.85(80) = Rs 68
To receive a 15 per cent discount, the shop must order at least 2000 films. Unfortunately,
the batteries’ shelf life is 18 months. The demand in 18 months is
78 × 21 = 1638 films
If he orders 2000 films, he would have to scrap 372 of them. This would cost more than
the 15 per cent discount would save.
(b) For C3 = 0.90(80) = Rs 72
Q3∗ = 2 DCo /Cc

= 2 (1092 )( 20 ) / ⎡⎣0.20 ( 72 ) ⎤⎦
= 55 films ( not feasiible)
The most economical, feasible quantity for C3 is 900 films. Since the carrying cost is
function of material cost and equals to Rs 72 only when order size is 900 films.
(c) For C2 = 0.93(80) = Rs 74.4
Q2∗ = 2 DCo /Cc

= 2 (1092 )( 20 ) / ⎡⎣0.20 ( 74.4 ) ⎤⎦

= 54.18 films ( not feasible)
The most economical, feasible quantity for C2 is 400 films. Since the carrying cost is a
function of material cost and equals to Rs 74.4 only when order size is 400 films.
(d) For C1 = 1.00(80) = Rs 80

Q1∗ = 2 DCo /Cc

= 2 (1092 )( 20 ) / 0.2 ( 80 )
= 52.24 films (feasible)
We know that
TCi = (1/ 2)(Qi∗Cc ) + ( DCo /Qi∗ ) + DCi
 Inventory Control  123

Therefore, we have
TC3 = (1/ 2)(900)(72) + ((1092)( 20) / (900)) + (1092)(14.4) = Rs 48,149.06
TC2 = (1/ 2)( 400)(74.4) + ((1092)( 20) / ( 400)) + (1092)(14.88) = Rs 31,183.56
TC1 = (1/ 2)(53)(80) + ((1092)( 20) / 53) + (1092)(16) = Rs 20, 004.07

Comparing the total costs for 53, 400 and 900 films, the lowest total annual cost is
Rs 20,004.07. The shop should order 53 films at a time.

6.6 Reorder Point

The reorder point is the inventory level at which a new order is placed. The order must be placed
while there is enough stock in place to cover demand during lead time
R = dL
where d is the demand rate per time period and L is the lead time. The inventory level might be
depleted at a slower or faster rate during lead time. When demand is uncertain, safety stock is
added as a hedge against stock-out.
The service level is the probability that the amount of inventory on hand is sufficient to
meet demand during lead time (i.e., the probability of a stock-out will not occur). The higher
the probability of existing inventory results in more likelihood of meeting the customer demand.
A service level of 95 per cent means there is a 0.95 probability that demand will be met and
0.5 is the probability of a stock-out.

6.6.1 Reorder Point for Variable Demand

There are three situations for determining the reorder point: reorder point with variable demand,
reorder point for variable lead time, and reorder point for both variable demand and lead time.
If the daily demand is a variable and lead time of supply is a constant, the following
expression is used to find the reorder point. The first term of the reorder point, i.e. dL represents
the average inventory required for the lead-time period and the second term, i.e. Z σ d L
represents the safety stock.

R = dL + Z σ d L

where R is the reorder point, d is the average daily demand, L is the lead time, σ d is the standard
deviation of daily demand, Z is the number of standard deviations corresponding to service level
probability, and Z σ d L is the safety stock.
In Figure 6.8, probability of meeting demand and probability of a stock-out are shown on
the standard normal distribution curve. The area lies in the left side of reorder point, R shows the
probability of meeting the demand and right side of the reorder point, R shows the probability of
a stock-out. The area between the mean value, i.e. dL and R shows the probability of meeting the
demand from the safety stock.
124  Industrial Engineering and Management

Probability of meeting demand

during lead time = service level

Probability of
a stock-out
Safety stock
Zσd√L

dL R
Demand

Figure 6-8: Probability distribution for stock-out

Example 6.5: A furniture house wants a reorder point with a 95 per cent service level and
a 5 per cent stock-out probability. The average demand per day is 75 chairs with a lead time
of 5 days. The standard deviation is 6 chairs per day. Find the reorder point.

Solution:
For a 95 per cent service level, Z = 1.65. So, the reorder point is

R = d × L + Zσ d L
= 75 (5) + (1.65)(6 ) (5 )
= 397.13 ≈ 398 chairs

6.6.2 Reorder Point for Variable Lead Time

When lead time is variable and daily demand is constant, following expression is used to
determine the reorder point:

R = d L + Zdσ L

where d is the constant daily demand, L is the average lead time, σ L is the standard deviation
of lead time, dσ L is the standard deviation of demand during lead time, and Zdσ L is the safety
stock.

Example 6.6: A cloth merchant has a fixed demand of 45 jackets per day. The average lead time
is 8 days and the deviation in lead time is 2 days. Calculate the reorder point with 95 per cent
service level.
 Inventory Control  125

Solution:
We have demand rate, d = 45 jackets per day, average lead time, L = 8 days, deviation in lead
time, σ = 2 days, Z corresponding to 95 per cent service level is 1.65 then,

R = d L + Zdσ
= 45 × 8 + 1.65 × 45 × 2
= 508.5
= 509 jackets

6.6.3 Reorder Point for Variable Demand and Variable Lead Time
When daily demand and lead time both are variables, the reorder point can be determined by the
following expression:
2
R = d L + Z (σ d ) 2 L + (σ L ) 2 d

(σ d ) L + (σ L ) d
2 2 2
where d is the average demand and L is the variable lead time, is the

(σ d ) L + (σ L ) d
2 2 2
standard deviation during the lead time Z is the safety stock.

Example 6.7: A cloth merchant has an average demand of 45 jackets per day with a standard
deviation of 6 jackets per day. The average lead time is 8 days and the deviation in lead time
is 2 days. Calculate the reorder point with 95 per cent service level.

Solution:
2
R = d L + Z (σ d ) 2 L + (σ L ) 2 d
= ( 45)(8) + (1.65) (6) 2 (8) + ( 2) 2 ( 45) 2
= 511.11 ≈ 512 jackets

6.7 Order Quantity for Variable Demand

In this chapter, we have studied to determine the order size, i.e. EOQ only when demand is
constant. Here, we will discuss the determination order size when demand is variable. Due to
variability in demand, we use safety stock in determination of order size as shown in following
expression:
Q = d (t b + L ) + Z σ d t b + L − I
where d is the average demand rate, tb is the fixed time between orders, L is the lead time, σ d is
the standard deviation of demand, Z σ d tb + L is the safety stock, and I is the inventory in stock.
126  Industrial Engineering and Management

Example 6.8:
A small shop has an average demand of 45 jackets per day and the standard deviation in average
demand is 6 jackets. The fixed time between orders is 16 days and the lead time is 8 days.
Inventory in stock is 40 jackets. For 95 per cent of service level what should be the order
quantity?

Solution:
Q = d (t b + L ) + Z σ d t b + L − I
= 45 (16 + 8) + 1.65 × 6 × 16 + 8 − 40
= 1088.49 jackets

 Summary
In this chapter, we have discussed about the importance of inventory and classification of
items in inventory based on their value, availability and consumption rate. In addition to
these discussions, various inventory models have been explained with numerical illustrations.
Also, reorder point has been explained with variable demand, variable lead time and variable
demand and lead time both. Finally, the order size for variable demand has been illustrated in
this text.

Multiple-Choice Questions
1. Which of the following does not belong to the assumption of economic batch quantity (EBQ)?
(a) Only two or more items are involved
(b) Annual demand is known
(c) Usage rate is constant
(d) Usage occurs continually
2. Inventory carrying costs are influenced by
(a) only ordering (b) only holding carrying cost per unit.
(c) both ordering and holding (d) only production cost
3. Which of the following statements concerning the basic EOQ model is true?
(a) a decrease in demand will increase the EOQ value
(b) the annual holding cost is less than the annual ordering cost for a smaller-order quantity compared
to EOQ
(c) an increase in holding cost will increase the EOQ value
(d) as annual ordering costs increase, so do annual carrying costs
4. ABC analysis deals with
(a) ordering cost (b) flow of materials
(c) ordering schedule of the materials (d) the cost involved with the materials
 Inventory Control  127

5. Which of the following are differences between periodic inventory and continuous inventory systems?
(a) Continuous inventory systems are time-phased while periodic inventory systems are not
(b) Periodic inventory systems have regular order times while continuous inventory systems have
irregular order times
(c) Periodic inventory systems have regular order quantities while continuous inventory systems have
irregular order quantities
(d) none of the above
6. A fixed time between orders and a variable order quantity are characteristics of a
(a) continuous inventory system (b) two-bin system
(c) periodic inventory system (d) MRP system
7. The function of inventory is
(a) to meet the fluctuating demand of the market
(b) to make smooth production systems
(c) to kelp hedge against price increase
(d) all of the above
8. VED analysis of inventory deals with
(a) utility of the materials (b) cost of the materials
(c) availability of the material (d) consumption of the material
9. FSN analysis of inventory deals with
(a) utility of the materials (b) cost of the materials
(c) availability of the material (d) consumption of the material
10. SDE analysis of inventory deals with
(a) utility of the materials (b) cost of the materials
(c) availability of the material (d) consumption of the material
11. Which of the following is not the part of assumptions of basic EOQ model?
(a) Demand is known with certainty and is relatively constant over time.
(b) Shortages are allowed.
(c) Lead time for the receipt of orders is constant.
(d) The entire order quantity is received at once and instantaneously.
12. Which of the following EOQ model assumes that material cost is the variable cost?
(a) Basic EOQ model with instantaneous receipt.
(b) EOQ model without instantaneous receipt.
(c) EOQ model with shortages.
(d) EOQ model with quantity discount.
13. To determine the reorder point, which of the following information is required?
(a) Demand rate (b) Lead time
(c) Safety stock (d) All the above
128  Industrial Engineering and Management

14. A firm has consumption rate of the material is 20 units per week and lead time of supply of the raw
material is two weeks. Which of the following will be the reorder point?
(a) 20 units (b) 40 units
(c) 60 units (d) 80 units
15. In the basic model of EOQ with instantaneous supply
(a) Ordering cost will be less than the holding cost.
(b) Ordering cost will be equal to the holding cost.
(c) Ordering cost will be more than the holding cost.
(d) There is no relationship between ordering cost and holding cost.

Answers
1. (a) 2. (c) 3. (b) 4. (d) 5. (b) 6. (c) 7. (d) 8. (a) 9. (d)
10. (c) 11. (b) 12. (d) 13. (d) 14. (b) 15. (b)

Review Questions
1. What are the various types of inventory maintained in a manufacturing organization?
2. What are the objectives of inventory?
3. Discuss the functions of inventory.
4. Write short notes on the following
(a) ABC analysis
(b) VED analysis
(c) FSN analysis
(d) SDE analysis
5. Derive an expression for economic order quantity subject to instantaneous supply, no shortage, and
without any discount.
6. Discuss the reorder point for variable demand and variable lead time.

Exercises
1. A company ABC is planning to stock a product A. The company has developed the following
information:
Annual usage = 6000 units
Cost of the product = Rs 350 /unit
Ordering cost = Rs 50 per order
Carrying cost = 25 per cent of the materials cost per unit per year.
(a) Determine the optimal number of units per order
(b) Find the optimal number of orders/year
(c) Find the annual total inventory cost
 Inventory Control  129

2. A furniture house decides to use the EOQ model to determine the order size of furniture from a
carpenter. The inventory manager observes that the annual demand of the furniture is 500 units for
last two years and expected to be same in the coming years. He has estimated that preparation of an
order and other variable costs associated with each order are about Rs 250, and it costs him about
20 per cent per year to hold items in stock. His cost for the furniture is Rs 1000.
(a) How many furniture should be ordered each time?
(b) How many orders would there be?
(c) Determine the approximate length of cycle time.
(d) Calculate the minimum total inventory cost.
3. An auto component manufacturer supplies headlight assembly to car dealers. The annual demand is
approximately 6000 headlight assemblies. The supplier pays Rs 2000 for each headlight assembly
and estimates that the annual holding cost is 15 per cent of the headlight assemblies’ value. It costs
approximately Rs 500 to place an order. The supplier currently orders 500 headlights per month.
(a) Determine the total inventory costs for the current order quantity.
(b) Determine the economic order quantity (EOQ).
(c) How many orders will be placed per year using the EOQ?
(d) Determine the total inventory costs for the EOQ.
4. In Exercise 3, upon closer inspection, the supplier determines that the demand for headlight assembly
is normally distributed with mean 4 headlight assemblies per day and standard deviation 3 headlight
assemblies per day. (The supplier plant is open 300 days per year.) It usually takes about 4 days to
receive an order from the factory.
(a) What is the standard deviation of usage during the lead time?
(b) Determine the reorder point needed to achieve a service level of 95 per cent.
(c) What is the safety stock? What is the holding cost associated with this safety stock?
(d) How would your analysis change if the service level changed to 98 per cent?

 References and Further Readings

1. Camp, W. E. (1922), ‘Determining the Production Order Quantity’, Management Engineering, 2:
17–18.
2. Hariga, M. and Ben Daya, M. (1999), ‘Some Stochastic Inventory Models with Deterministic Variable
Lead Time’, European Journal of Operational Research, 113: 42–51.
3. Harris, F. W. (1913). ‘How Many Parts to Make at Once. Factory’, The Magazine of Management,
10(2): 135–136. [Reprinted in 1990, Management Science, 38, 947–950.]
4. Khouja. M. and Sungjune P. (2003), ‘Optimal Lot Sizing Under Continuous Price Decrease’, Omega,
December 2003.
5. Lan, S. P., Chu, P., Chung, K. J., and Wan, W. J. (1999), ‘A Simple Method to Locate the Optimal
Solution of Inventory Model with Variable Lead Time’, Computer and Operations Research, 26:
599–605.
130  Industrial Engineering and Management

6. Magson, D. (1979), ‘Stock Control When the Lead Time Cannot be Considered Constant’, Journal
of Operational Research Society, 30: 317–322.
7. Naddor, E. (1966), Inventory Systems (New York: Wiley).
8. Ouyang, L. Y. and Chuang, B. R. (2000), ‘A Periodic Review Inventory Model Involving Variable Lead
Time with a Service Level Constraint’, International Journal of Systems Science, 31: 1209–1215.
9. Taft, E. W. (1918), ‘The Most Economical Production Lot’, Iron Age, 101: 1410–1412.
10. Wemmerlov, U. (1982), ‘Inventory Management and Control’, in Ed. G. Salvendy (ed), Handbook of
Industrial Engineering (New York, John Wiley and Sons).
11. Wilson, R. H. (1934). ‘A Scientific Routine for Stock Control’, Harvard Business Review, 13: 116–128.