Rolling Pem m c02 Forecasting

PEM M C02: Operations ManagementDr. ANAND PRAKASHPost Graduate Programme in Project Engineering and Management

Forecasting

What is Forecasting?Process of predicting a future eventUnderlying basis of all business decisionsProductionInventoryPersonnelFacilities

Short-range forecastUp to 1 year, generally less than 3 monthsPurchasing, job scheduling, workforce levels, job assignments, production levelsMedium-range forecast3 months to 3 yearsSales and production planning, budgetingLong-range forecast3+ yearsNew product planning, facility location, research and developmentForecasting Time Horizons

Types of ForecastsEconomic forecastsAddress business cycle inflation rate, money supply, housing starts, etc.Technological forecastsPredict rate of technological progressImpacts development of new productsDemand forecastsPredict sales of existing products and services

Forecasting ApproachesUsed when situation is vague and little data existNew productsNew technologyInvolves intuition, experiencee.g., forecasting sales on InternetQualitative Methods

Forecasting ApproachesUsed when situation is stable and historical data existExisting productsCurrent technologyInvolves mathematical techniquese.g., forecasting sales of color televisionsQuantitative Methods

Overview of Quantitative ApproachesNaive approachMoving averagesExponential smoothingTrend projectionLinear regression

Set of evenly spaced numerical dataObtained by observing response variable at regular time periodsForecast based only on past values, no other variables importantAssumes that factors influencing past and present will continue influence in futureTime Series Forecasting

Time Series Components

Components of DemandAverage demand over 4 yearsTrend componentRandom variation

Persistent, overall upward or downward patternChanges due to population, technology, age, culture, etc.Typically several years duration Trend Component

Regular pattern of up and down fluctuationsDue to weather, customs, etc.Occurs within a single year Seasonal Component

Repeating up and down movementsAffected by business cycle, political, and economic factorsMultiple years durationOften causal or associative relationshipsCyclical Component

Erratic, unsystematic, residual fluctuationsDue to random variation or unforeseen eventsShort duration and non-repeating Random Component

MA is a series of arithmetic means Used if little or no trendUsed often for smoothingProvides overall impression of data over timeMoving Average Method

Form of weighted moving averageWeights decline exponentiallyMost recent data weighted mostRequires smoothing constant ()Ranges from 0 to 1Subjectively chosenInvolves little record keeping of past dataExponential Smoothing

Exponential SmoothingNew forecast =Last periods forecast+ a (Last periods actual demand Last periods forecast)Ft = Ft 1 + a(At 1 - Ft 1)whereFt=new forecastFt 1=previous forecasta=smoothing (or weighting) constant (0 a 1)

Common Measures of Error

Common Measures of Error

Trend ProjectionsFitting a trend line to historical data points to project into the medium to long-rangeLinear trends can be found using the least squares technique

Seasonal Variations In DataThe multiplicative seasonal model can adjust trend data for seasonal variations in demand

Associative ForecastingUsed when changes in one or more independent variables can be used to predict the changes in the dependent variableMost common technique is linear regression analysisWe apply this technique just as we did in the time series example

Associative ForecastingForecasting an outcome based on predictor variables using the least squares technique

Associative Forecasting Example

Associative Forecasting Example

Associative Forecasting ExampleSales = 1.75 + .25(payroll)If payroll next year is estimated to be $6 billion, then:Sales = 1.75 + .25(6)Sales = $3,250,000

Standard Error of the EstimateA forecast is just a point estimate of a future value

This point is actually the mean of a probability distribution

Standard Error of the Estimatewherey=y-value of each data pointyc=computed value of the dependent variable, from the regression equationn=number of data points

How strong is the linear relationship between the variables?Correlation does not necessarily imply causality!Coefficient of correlation, r, measures degree of associationValues range from -1 to +1Correlation

Correlation Coefficient

r =

Coefficient of Determination, r2, measures the percent of change in y predicted by the change in xValues range from 0 to 1Easy to interpretCorrelation

Multiple Regression AnalysisIf more than one independent variable is to be used in the model, linear regression can be extended to multiple regression to accommodate several independent variablesComputationally, this is quite complex and generally done on the computer

***********************