Title: Forecasting for Operations
1Forecasting for Operations
2Forecasting for operations
- Research themes
- The damped trend
- Case studies
- Supply chain costs Specialty chemicals
- Manufacturing inventory investment Snack foods
- Purchasing workload Water treatment systems
- Consequences of forecast errors
- How to evaluate forecast performance
3Research themes
- Intermittent demand
- Distribution inventory management
- Biased forecasting
- Bullwhip effect
- Sensitivity of costs to forecast errors
4Intermittent demand
- Empirical research is mixed - not clear that
intermittent methods can beat SES - No underlying model exists for the Croston method
or any of its variants (Shenstone Hyndman, IJF,
2005) - Why not remove zeroes by aggregation?
(Nikolopoulos et al.,JORS, 2011)
5Distribution inventory management
- The damped trend gives better inventory
performance than other exponential smoothing
methods (Gardner, MS, 1990) - Marginal improvements in forecast accuracy
produce much larger improvements in inventory
costs (Syntetos et al., IJF, 2010)
6Biased forecasting
- Effects (Sanders Graman, Omega,2009)
- Costs are more sensitive to bias than variance
- Over-forecasting produces lower costs than
unbiased forecasting in an MRP environment - Objections
- Conclusions depend on assumptions
- Safety stock is always a better option than
adding bias to the forecasts
7The bullwhip effect
- Definition
- Tendency of demand variability to increase as one
moves up a supply chain - Caused by lead times and forecast errors
- Is the bullwhip effect inevitable?
- Yes But it can be reduced with centralized
demand information (Chen et al., MS, 2000) - No Bullwhip effect is due to poor research
design (Fildes Kingsman, JORS, 2010)
8Sensitivity of costs to forecast error
- Fildes and Kingsman (JORS, 2011)
- Research design
- MRP simulation
- Distinguishes between noise and specification
error - Demand processes are experimental factors
- Conclusions
- Cost increases exponentially with demand
uncertainty - Cost benefits of improved forecasting are greater
than the effects of choosing inventory decision
rules
9Performance of the damped trend
- The damped trend is a well established
forecasting method that should improve accuracy
in practical applications. (Armstrong, IJF,
2006) - The damped trend can reasonably claim to
be a benchmark forecasting method for
all others to beat. (Fildes et al., JORS, 2008)
10Why the damped trend works
- Rationale
- The damped trend has an underlying random
coefficient state space (RCSS) model that adapts
to changes in trend (McKenzie Gardner, IJF,
2011) - Practice
- Fitting the damped trend is a means of automatic
method selection from numerous special cases
(Gardner McKenzie, JORS, 2011)
11SSOE state space models
- At are i.i.d. binary random variates
- White noise innovation processes e and are
different - Parameters h and h are related but usually
different
12Runs of linear trends in the RCSS model
- With a strong trend, At will consist of long
runs of 1s with occasional 0s.
- With a weak trend, At will consist of long
runs of 0s with occasional 1s. - In between, we get a mixture of models on shorter
time scales, i.e. damping.
13Advantages of the RCSS model
- Allows both smooth and sudden changes in trend.
- is a measure of the persistence of the linear
trend. The mean run length is thus -
- RCSS prediction intervals are much wider than
those of constant coefficient models.
and
14Methods automatically identified in the M3 time
series
Method
Damped trend 43.0
Holt 10.0
SES w/ damped drift 24.8
SES w/ drift 2.4
SES 0.8
RW w/ damped drift 7.8
RW w/ drift 2.5
RW 0.0
Modified exp. trend 8.3
Linear trend 0.1
Simple average 0.3
15Case 1 Chemicals supply chain
- Scope
- 4 plants N. and S. America, Europe, Asia
- 10 component chemicals, 25 products
- 400 customers, 250,000 tons of annual production
- Production and transportation plans based on
- Damped trend
- Optimization
- Simulation
16Examples of chemicals demand series
1
2
3
4
17Scaled errors
- Average forecast error measures are misleading
- Drastic changes in scale
- Some observations near zero
- Alternative - Scaled errors (Hyndman Koehler,
2006) - Based on in-sample, one-step errors from the
naïve method - If scaled error is less than 1, we beat the naïve
method
18(No Transcript)
19Proportions of total demand for 25 time series
20(No Transcript)
21Supply chain model
Damped trend
Actual demand
Simulation daily mfg. shipments
Monthly demand forecasts
MIP Minimize total supply chain cost
Inv. on hand Inv. in transit Backorders
Monthly production schedule
MIP Disaggregate monthly schedule
Detailed weekly schedule
22Top-level mixed integer program (MIP)
- Objective Minimize total supply chain costs,
including - Inventory carrying
- Production
- Transportation
- Import tariffs
23Top-level MIP continued
- Data requirements
- Demand forecasts
- Pending orders
- Shipments in transit
- Inventory levels
- Machine and storage capacity
- Business rules for
- Production run lengths
- Transportation modes
24Supply chain model
Damped trend
Actual demand
Simulation daily mfg. shipments
Monthly demand forecasts
MIP Minimize total supply chain cost
Inv. on hand Inv. in transit Backorders
Monthly production schedule
MIP Disaggregate monthly schedule
Detailed weekly schedule
25Second-level MIP
- Disaggregates top-level schedule
- Weekly schedule for each machine at each plant
- 12-week horizon
- Data requirements
- Forecasts
- Week-ending inventories
- Pending orders
- Scheduled in and out bound shipments
- Bootstrap safety stocks (Snyder et al., IJF,
2002)
26Supply chain model
Damped trend
Actual demand
Simulation daily mfg. shipments
Monthly demand forecasts
MIP Minimize total supply chain cost
Inv. on hand Inv. in transit Backorders
Monthly production schedule
MIP Disaggregate monthly schedule
Detailed weekly schedule
27Simulation model
- Executes manufacturing plans on a daily basis
using actual demand history - Feeds production, inventories, backorders, and
shipments to the MIP models - Sources of uncertainty
- Demand
- Transportation lead times
- Machine breakdowns
28(No Transcript)
29(No Transcript)
30Case 2 Snack-food manufacturer
- Scope
- 82 snack foods
- Food stocks managed by commodity traders
- Packaging materials managed with subjective
forecasts and EOQ/safety stock inventory rules - Problems
- Excess stocks of perishable packaging materials
- Difficult to predict inventory on the balance
sheet
3111-Oz. corn chipsMonthly packaging inventory and
usage
Actual Inventory from subjective forecasts
Month
Monthly Usage
32Snack-food manufacturer
- Solution
- Automatic forecasting with the damped trend
- Retain EOQ/safety stock inventory rules
33Damped-trend performance
11-oz. corn chips
Outlier
34Investment analysis 11-oz. corn chips
35Safety stocks vs. shortages
11-oz. corn chips
36Safety stock vs. forecast errors
11-oz. corn chips
Safety stock
Forecast errors
3711-Oz. corn chipsTarget vs. actual packaging
inventory
Actual Inventory from subjective forecasts
Actual Inventory from subjective forecasts
Target maximum inventory based on damped trend
Month
Monthly Usage
38Forecasting regional demand
- Forecast total unit demand with the damped trend
- Forecast regional percentages with simple
exponential smoothing
39Regional sales percentages Corn chips
40(No Transcript)
41Case 3 Water treatment company
- Scope
- Assembly of systems and distribution of supplies
- Annual sales 16 million
- Inventory 6 million (23,000 SKUs)
- Inventory system
- Reorder monthly to maintain 3 months of stock
- Numerous subjective adjustments
- Forecasting system
- 6-month weighted moving average
- Numerous subjective adjustments
42Problems
- Forecasts vs. reality
- Annual forecasts on stock records 29 million
- Annual sales 16 million
- Purchasing workload
- 76,000 purchase orders per year
- Messy stock records
- Dead stock
- Substitute items not linked to primary items
43Water treatment company
Inventory status
44Solutions
- Forecast demand with the damped trend
- Develop a decision rule for what to stock
- Use the forecasts to do an ABC classification
- Replace the monthly ordering policy with a hybrid
inventory control system - Class A JIT
- Class B EOQ/safety stock
- Class C Annual buys
45(No Transcript)
46What to stock?
- Cost to stock
- Average inventory balance x holding rate
- Number of stock orders x transportation cost
- Cost to not stock
- Nbr. of customer orders x drop-ship
transportation cost -
47 ABC classification based ondamped-trend
forecasts
Class Sales forecast System Items Dollars
A gt 36,000 JIT 3 75
B 600 - 35,999 EOQ 49 18
C lt 600 Annual buy 48 7
48Annual purchasing workload Total savings 58,000
orders (76)
EOQ
JIT
49Inventory investment Total savings 591,000
(15)
EOQ
JIT
50Consequences of forecast errors
- Limited capacity creates interactions amongst
products - Under-forecasting
- Chain reaction of backorders
- Premium transportation
- Over-forecasting
- Excess stocks
- Chain reaction of backorders (limited capacity
put to wrong use) - Premium transportation
51Consequences of forecast errors (cont.)
- Errors often reverse themselves before system has
fully responded to - Backorders, or
- Excess stocks
52How to evaluate forecast performance
- Operational measures
- Backorder delay time
- Percent of time in stock
- Percent of orders filled immediately
- Number of purchase orders or production setups
- Financial measures
- Manufacturing, distribution, and supply chain
costs - Value of backorders
- Inventory investment on the balance sheet
53Future research
- Research is needed
- In real operating systems
- Gardner Makridakis (IJF,1988)
- On the benefits of improved forecasting
- Fildes Kingsman (JORS, 2010)
- On the relationship between forecast accuracy and
operational performance - Syntetos et al. (IJF, 2010)
54 Presentation and papers available at
www.bauer.uh.edu/gardner