Chapter Two Materials Demand Forecasting

INTRODUCTION

Forecasting is important because it helps reduce uncertainty. Forecasting is the art and science of predicting future events. It involves estimation of the occurrence, timing, and/or magnitude of uncertain future events or levels of activities.

FEATURES COMMON TO ALL FORECASTS

Assume that the same underlying causal system that existed in the past will continue to exist in the future Rarely perfect Forecasts for groups of items tend to be more accurate than individual items Accuracy as the time period covered by the forecast

THE FORECASTING PROCESS

1.

2.
3. 4.

5.
6.

Determine the purpose of the forecast Determine the time horizon Select an appropriate technique Identify the necessary data and gather it Make the forecast Monitor forecast errors

APPROACHES TO FORECASTING

There are two general approaches to forecasting: Qualitative methods consists mainly subjective inputs, based on judgment about the casual factors. Quantitative methods involve either the extension of historical data or the development of associative models

QUALITATIVE METHODS

Individual Opinion: consists of collecting opinions and judgments of individuals Executive Committee Consensus: a committee of executives is constituted with the responsibility of developing a forecast. The Delphi Method: seeks to eliminate the undesirable consequences of group thinking Field Expectation Method: individual members of sales force are required to submit sales forecasts of their respective regions. Users Expectation method: estimates of future sales are obtained directly from customers. Historical Analogy: estimate of future sales of product to knowledge of a similar products sales. Market Surveys: questionnaires, telephone talks or field

TIME SERIES FORECASTING

Time series is a collection of data of some economic variable or composite of variables recorded over a period time- weekly, monthly, quarterly, or yearly. It is an arrangement of statistical data in chronological order. Time-series models predict on the assumption that the future is a function of the past.

DECOMPOSITION OF A TIME SERIES

The Secular Trend (T) is the smooth long-term direction of a time series. The Cyclical Variation (C) is the rise and fall of a time series over periods longer than one year The Seasonal Variation (S): are patterns of change in a time series with in a year The Irregular Variation is a time series is subjected to occasional influences

FIGURE 1

EXAMPLES OF PATTERNS IN TIME SERIES DATA

NAIVE FORECAST

assume that the forecast in the next period will be equal to demand in the most recent period may take in to account a demand trend

MOVING AVERAGE

Attempts to forecast values on the basis of the average of the values of past few periods. The formula for computing the simple moving average is as follows:

MAn

Di
i 1

EXAMPLE

The Bright Company sells and delivers office supplies to various companies, schools, and agencies. The office supply business is extremely competitive, and the ability to deliver orders promptly is an important factor in getting new customers and keeping old ones. The manager of the company wants to be certain that enough drivers and delivery vehicles are available so that orders can be delivered promptly. Therefore, the manager wants to be able to forecast the number of orders that will occur during the next month. From records of delivery orders, the manager has accumulated data for the past 10 months. These data are shown below:

3- AND 5-MONTH AVERAGES

FIGURE
140 120
Quantity Ordered

100 80 60 40 20 0 1 2 3 4 5 6 Month 7 8 9 10 11 Order Per Month 3 Month Moving Average 5 Month Moving Average

WEIGHTED MOVING AVERAGE

Involves making a forecast values by giving differential weights to the values entering into moving average calculation.
WMA n Wi Di
i 1 n

Example: If the Bright Company wants to compute a 3-month weighted moving average with a weight of 50% for the October data, a weight of 30% for the September data, and a weight of 20% for August, it is computed as:

WMA 3 WiDi (. 50 )(90 ) (.30 )(110 ) (. 20 )(130 ) 104

i 1

EXPONENTIAL SMOOTHING

the forecast for the next period is calculated as weighted average of all the previous values. We will consider two forms of exponential smoothing: simple exponential smoothing and adjusted exponential smoothing The simple exponential smoothing forecast is computed by using the formula:

Ft 1 Dt (1 ) Ft

EXAMPLE
PM Computer Services assembles customized personal computers from generic parts. The company was formed and is operated by two part-time Hawassa University students, Mark and John, and has had steady growth since it started. The company assembles computers mostly at night, using other part-time students as labor. Mark and John purchase generic computer parts in volume at a discount from a variety of sources whenever they see a good deal. It is therefore important that they develop a good forecast of demand for their computers so that they will know how many computer component parts to purchase and stock. The company has accumulated the demand data for its computers for the past 12 months, from which it wants to compute exponential smoothing forecasts, using smoothing constants equal to .30 and .50.

Exponential Smoothing Forecasts, = .30 AND = .50

EXERCISE

KK Textile Factory uses cotton as a raw material in order to produce various textile products. Assume that an initial starting forecasted value of 1,000 tons for year 2012 and of 0.2. If actual demand turned to out to be 1,100 tons rather than the forecast, what is the forecast for the year 2013?

Adjusted Exponential Smoothing

Consists of the exponential smoothing forecast with a trend adjustment factor added to it. The formula for the adjusted forecast is: AFt+1 = Ft+1 + Tt+1 A forecast model for trend: Tt+1 = (Ft+1 - Ft) + (1 - )Tt

EXAMPLE

PM Computer Services now wants to develop an adjusted exponentially smoothed forecast, using the same 12 months of demand. It will use the exponentially smoothed forecast with = .50 computed in with a smoothing constant for trend of .30. The adjusted forecast for March: T3 = (F3 - F2) + (1 - )T2 = (.30)(38.5 37.0) + (.70)(0) = 0.45 AF3 = F3 + T3 = 38.5 + 0.45 = 38.95

Adjusted Exponentially Smoothed Forecast

LINEAR TREND LINE

A linear trend line relates a dependent variable, which for our purposes is demand, to one independent variable, time, in the form of a linear equation, as follows:

Y = a + bx

These parameters of the linear trend line can be calculated by using the least squares formulas for linear regression: x xy nx y x b n 2 2 x nx y y a y bx n

LEAST SQUARES CALCULATION

Therefore, the linear trend line is Y = 35.2 + 1.72X To calculate a forecast for period 13, X = 13 would be substituted in the linear trend line: Y = 35.2 + 1.72(13) = 57.56

EXERCISE

1) 2)

A cosmetics manufacturer production department has developed a linear trend equation that can be used to predict the annual material requirement used to produce its popular product. Y = 80 + 15X Where: Y = Annual materials requirement (000 units) X = 1993 Are annual materials requirement increasing or decreasing? By how much? Predict annual materials requirement for 2005 using the equation

SEASONAL ADJUSTMENTS

A SEASONAL FACTOR is a numerical value that is multiplied by the normal forecast to get a seasonally adjusted forecast.

Di Si D
Example: ELFORA Farms is a company that raises sheep, which it sells to a meat-processing company throughout the year. However, the peak season obviously occurs during the fourth quarter of the year, October to December. ELFORA Farms has experienced a demand for sheep for the past 3 years as shown

Y = 40.97 + 4.30X= 40.97 + 4.30(4) = 58.17

Using this annual forecast of demand, the seasonally adjusted forecasts for 2006 are as follows: SF1 = (S1) (F5) = (.28) (58.17) = 16.28 SF2 = (S2) (F5) = (.20) (58.17) = 11.63 SF3 = (S3) (F5) = (.15) (58.17) = 8.73 SF4 = (S4) (F5) = (.37) (58.17) = 21.53

CAUSAL RELATIONSHIP FORECASTING

Simple Linear Regression is a method of forecasting similar to time series analysis. Time series analysis tries to see the relationship between a variable and time, where as regression analysis observes relationship existing between two variables not necessarily time. Example: Suppose you are a general manager of a building material manufacturing plant, and you feel that the demand for the brick that your company selling is related to the construction permit issued by the municipality where you sell the product.

You have collected the following data.

Calculate the demand for the bricks if you discover that the municipality has issued 30 permits for the next year

CORRELATIONS

Correlation measures the strength and direction of relationship between two variables. Correlation can range from 1.00 to + 1.00. A correlation close to zero indicates little linear relationship between two variables. The correlations between two variables can be computed using the equation below:

EXERCISE

The owner of a small hardware store has noted a sales pattern for window locks that seems to parallel the number of break-ins reported each week in the newspaper. The data are:

1. Develop a regression equation for the data 2. Estimate sales when the number of break-ins is five

FORECAST ACCURACY The difference between the forecast and the actual is referred to as the forecast error The objective of forecasting is for the error to be as slight as possible. There are a variety of different measures of forecast error: Mean Absolute Deviation (MAD) Mean Absolute Percent Deviation (MAPD) Cumulative Error (E) Average Error or Bias Mean Squared Error (MSE)

Mean Absolute Deviation (MAD)

MAD is an average of the difference between the forecast and actual demand
/D MAD
t

Ft /

We cannot compare a MAD value of two different data and say the former is good and the latter is bad; they depend to a certain extent on the relative magnitude of the data. MAD.doc

Mean Absolute Percent Deviation (MAPD)

MAPD measures the absolute error as a percentage of demand rather than per period.

/D F / MPAD D
t t t

Cumulative Error (E)

Cumulative Error is computed simply by summing the forecast errors

E et

AVERAGE ERROR

Computed by averaging the cumulative error over the number of time periods

e E n

MEAN SQUARED ERROR

Each individual error value is squared, and then these values are summed and averaged.

McGraw -

EXERCISE

A computer software firm has experienced the following demand for its Personal Finance software package:

Develop an exponential smoothing forecast, using = .40, and an adjusted exponential smoothing forecast, using = .40 and = .20. Compare the accuracy of the two forecasts, using MAD and cumulative error(assume the forecast of the first period is the same with the actual demand and the trend is zero.

End of Chapter Two