Supply Chain Management Lecture 11

Outline Today –Homework 2 –Chapter 7 Thursday –Chapter 7 Friday –Homework 3 Due Friday February 26 before 5:00pm

Announcements FEI Student Financial Awards Program –Awards are presented to Finance or Accounting majors from schools in Colorado. Each award can either go to an undergraduate or graduate student. This year there are five awards for $1,200 each. –The criteria includes the following three factors: Students who have performed well academically, Students who have potential leadership skills in the business field, and Students who have financial need. –All applications are due to the committee no later than March 25, Applications and information are available in the office of Bonnie Beverly (KOBL S315A) or Consuelo Delval (KOBL S328) (paper applications only)

Announcements What? –Tour the Staples Fulfillment Center in Brighton, CO –Informal Lunch-and-Learn –Up to 20 students with a Operations Management major When? –Weeks of March 15 or March 29 –There is a fair amount of time involved in the activity Transit is close to an hour in each direction Probably 2 hours onsite

Forecasting Examples Walt Disney World –Daily forecast of attendance (weather forecasts, previous day’s crowds, conventions, seasonal variations) –Add more cast members and add street activities to manage high demand Amazon Kindle –Kindle sold out in 5.5 hours –Kindle was not in stock for another 5 months FedEx customer service center –Goal is to answer 90% of all calls within 20 seconds –Makes extensive use of forecasting for staffing decisions and to ensure that customer satisfaction stays high

Characteristics of Forecasts 1.Forecasts are always wrong! 2.Long-term forecasts are less accurate than short-term forecasts 3.Aggregate forecasts are more accurate than disaggregate forecasts 4.Information gets distorted when moving away from the customer

Types of Forecasts Qualitative –Primarily subjective, rely on judgment and opinion Time series –Use historical demand only Causal –Use the relationship between demand and some other factor to develop forecast Simulation –Imitate consumer choices that give rise to demand

Role of Forecasting Push Pull ManufacturerDistributorRetailerCustomerSupplier Is demand forecasting more important for a push or pull system?

Time Series Forecasting Observed demand = Systematic component + Random component LLevel (current deseasonalized demand) TTrend (growth or decline in demand) SSeasonality (predictable seasonal fluctuation) The goal of any forecasting method is to predict the systematic component of demand and estimate the random component

Components of an Observation Level (L)Forecast(F) F t+n = L t The moving-average method is used when demand has no observable trend or seasonality

Example: Moving Average Method A supermarket has experienced the following weekly demand of coffee over the last four weeks –120, 127, 114, and 122 Determine Level L t = (D t +D t-1 +…+D t-N+1 )/N Forecast F t+n = L t

Example: Tahoe Salt

Demand forecasting using Moving Average

Components of an Observation Level (L)Forecast(F) F t+n = L t The simple exponential smoothing is used when demand has no observable trend or seasonality

Example: Simple Exponential Smoothing Method A supermarket has experienced the following weekly demand of coffee over the last four weeks –120, 127, 114, and 122 Determine initial level L 0 = (∑ i D i )/ n Determine levels L t+1 =  D t+1 + (1 –  )*L t Forecast F t+n = L t  = 0.1

Example: Tahoe Salt

Demand forecasting using simple exponential smoothing

Components of an Observation Trend (T) Forecast(F) F t+n = L t + nT t Holt’s method is appropriate when demand is assumed to have a level and a trend

Example: Holt’s Method An electronics manufacturer has seen demand for its latest MP3 player increase over the last six months –8415, 8732, 9014, 9808, 10413, Determine initial level L 0 = INTERCEPT(y’s, x’s) T 0 = LINEST(y’s, x’s)

Example: Holt’s Method An electronics manufacturer has seen demand for its latest MP3 player increase over the last six months –8415, 8732, 9014, 9808, 10413, Determine initial level L 0 = INTERCEPT(y’s, x’s) T 0 = LINEST(y’s, x’s) Determine levels L t+1 =  D t+1 + (1 –  )*(L t + T t ) T t+1 =  (L t+1 – L t ) + (1 –  )*T t Forecast F t+n = L t + nT t  = 0.1,  = 0.2

Example: Tahoe Salt

Demand forecasting using Holt’s method

Components of an Observation Seasonality (S) Forecast(F) F t+n = (L t + T t )S t+n

Time Series Forecasting Observed demand = Systematic component + Random component LLevel (current deseasonalized demand) TTrend (growth or decline in demand) SSeasonality (predictable seasonal fluctuation)

Static Versus Adaptive Forecasting Methods Static –D t : Actual demand –L: Level –T: Trend –S: Seasonal factor –F t : Forecast Adaptive –D t : Actual demand –L t : Level –T t : Trend –S t : Seasonal factor –F t : Forecast

Example: Static Method A theme park has seen the following attendance over the last eight quarters (in thousands) –54, 87, 192, 130, 80, 124, 265, 171 Determine initial level L = INTERCEPT(y’s, x’s) T = LINEST(y’s, x’s) Determine deason. demand D t = L + Tt Determine seasonal factors S t = D t / D t Determine seasonal factors S i =AVG(S i ) Forecast F t = (L + T t )S i

Example: Tahoe Salt

Static Forecasting Method

Deseasonalize demand –Demand that would have been observed in the absence of seasonal fluctuations Periodicity p –The number of periods after which the seasonal cycle repeats itself 12 months in a year 7 days in a week 4 quarters in a year 3 months in a quarter

Deseasonalize demand

Periodicity p is oddPeriodicity p is even

Deseasonalizing demand around t= (2,4), that is, year 2 and 4th quarter, when p is odd Deseasonalize demand

Assume p = 3, hence a seasonal cycle consists of three periods

Deseasonalize demand Deseasonalized demand for t=(2,4) = 18, , ,000 = 26,333

Deseasonalize demand Deseasonalizing demand around t= (2,4), that is, year 2 and 4th quarter, when p is even

Deseasonalize demand Assume p = 4, hence a seasonal cycle consists of four periods

Deseasonalize demand What happens if you take the average demand?

Deseasonalize demand

Periodicity p is oddPeriodicity p is even

Example: Tahoe Salt

Static Forecasting Method

Determine initial level L = INTERCEPT(y’s, x’s) T = LINEST(y’s, x’s) Determine deason. demand D t = L + Tt Determine seasonal factors S t = D t / D t Determine seasonal factors S i =AVG(S i ) Forecast F t = (L + T t )S i Deasonalize demand Depends on number periods in a seasonal cycle

Example: Tahoe Salt Demand forecast using Static forecasting method

Example: Winter’s Model A theme park has seen the following attendance over the last eight quarters (in thousands) –54, 87, 192, 130, 80, 124, 265, 171 Determine initial levels L 0 = From static forecast T 0 = From static forecast S i,0 = From static forecast Forecast F t+1 = (L t + T t )S t+1 Determine levels L t+1 =  (D t+1 /S t+1 )+ (1 –  )*(L t + T t ) T t+1 =  (L t+1 – L t ) + (1 –  )*T t S t+p+1 =  (D t+1 /L t+1 ) + (1 –  )*S t+1

Example: Tahoe Salt

Demand forecast using Winter’s method