Supply Chain Management

Lecture 14
Outline
• February 25 (Today)
– Network design simulation description
– Chapter 8
– Homework 4 (short)
• March 2
– Chapter 8, 9
– Network design simulation due before 5:00pm
• March 4
– Simulation results
– Midterm overview
– Homework 4 due
• March 9
– Midterm
Measures of Forecast Error
Error measure Formula

Et Forecast Error Ft - Dt

Biast Bias ∑tn=1 Et

At Absolute Deviation |Et|

MADn Mean absolute deviation (1/n)*∑nt=1 At

TSt Tracking signal Biast / MADt

MAPEn Mean absolute percentage (1/n)*∑nt=1 (At / Dt)*100

error
MSEn Mean squared error (1/n)*∑nt=1 Et2
Measures of Forecast Error
Error measure Description

Et Forecast Error Forecast – Demand

Biast Bias Sum of errors

At Absolute Deviation Absolute error

MADt Mean absolute deviation Average of absolute error

TSt Tracking signal Biast / MADt

MAPEn Mean absolute percentage Average of absolute percentage

error error
MSEn Mean squared error Average of squared error
Measures of Forecast Error
Error measure Desired outcome / Use

Et Forecast Error Close to zero

Biast Bias Close to zero

At Absolute Deviation Close to zero

MADt Mean absolute deviation STDEV(Et)  1.25 MADt

TSt Tracking signal Stay within (-6, +6)

MAPEn Mean absolute percentage Stay under 10% (30% not

error uncommon)
MSEn Mean squared error VAR(Et)  MSEt
Simulation Assignment (25%)
• Design the supply chain network for Jacobs Industries on
the fictional continent of Pangea
– Jacobs only product is an industrial chemical that can be mixed
with air to form a foam (used in air conditioner retrofit kits)
Demand

• Demand for Jacob’s product in Pangea

– Existing and new markets

Hardwood floor
Air conditioner
laminates
retrofit kit

Premium home
Premium home Insulation appliances
appliances products
Demand

• Average demand for Jacob’s product in Pangea

– Existing and new markets

140

120

120 100

100 80

60
80
40
60
20
40
0
20 1 145 289 433 577 721 865 1009 1153 1297 1441

0
1 145 289 433 577 721 865 1009 1153 1297 1441

18
16
14

18 12

16 10
8
14
18 6
12
16

250
4
10
14 2
8
12 0
6
10 1 142 283 424 565 706 847 988 1129 1270 1411
4
8
2
6
0
4
1 142 283 424 565 706 847 988 1129 1270 1411
2
0
1 142 283 424 565 706 847 988 1129 1270 1411
Assignment
• Jacobs management would like to design a supply chain
network for Pangea. It’s current network consist of a
factory in Calopeia with a capacity of 20. You have been
hired to suggest a network design that will maximize
profits for Jacobs Industry. Designing such a network is
complex and includes the following decisions:
– Should the factory in Calopeia be expanded?
– Should factories in other regions be built? If so, what should their
capacity be?
– What regions should each factory serve?
Serve region?
Total
Factory? Capacity? Calopeia Sorange Tyran Entworpe Fardo
Calopeia YES 40 YES YES YES YES YES
Sorange
Tyran
Entworpe YES 20 YES YES YES YES NO
Fardo
Production parameters

20
Production parameters

• A factory must serve the region in which it is

located

300
Production parameters

20
20

20 20
20
Production parameters

• You have $20,000,000 to design your network

• The cost of building a factory is $500,000
regardless of the factory capacity
• The cost of capacity is $50,000

20

500,000 + 5*50,000 =
5 $750,000
Production parameters

• You have $20,000,000 to design your network

• The cost of building a factory is $500,000
regardless of the factory capacity
• The cost of capacity is $50,000

20*50,000 =
40 $1,000,000

500,000 + 5*50,000 =
5 $750,000
Transportation parameters

• Finished drums are shipped from the factory

warehouse by mail to the customers
• Factories may ship to all the regions in Pangea
• Shipping time is 1 day independent of origin and
destination
To Calopeia To Sorange To Tyran To Entworpe To Fardo
From Calopeia 50 100 100 100 200
From Sorange 100 50 100 100 200
From Tyran 100 100 50 100 200
From Entworpe 100 100 100 50 200
From Fardo 200 200 200 200 50
Financial and Other Parameters
• All customers pay $1450 per drum and the production cost
is $1200 per drum
• The drum must be shipped within 24 hours of receiving
the order or the order is lost
• Orders may be partially filled and one order may be filled
from multiple factories
• Each factory has warehouse space to hold up to 500
finished drums
– If warehouse space is used completely, the factory will remain idle
until warehouse space becomes available
• Interest accrues on cash at 10% per year, compounded
daily
The Goal

• Your network design will run from day 1 till day

1460
• Investment in capital (such as new factories and
factory capacity) will become obsolete on day
1460

The winning team is the one with the

highest cash position on day 1460
From Forecasting to Planning

2500
Forecast
2000
Demand

1500
Capacity
1000

500

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Month
How should a company best utilize the
resources that it has?
Aggregate Planning Strategies
• Basic strategies
– Level strategy (using inventory as lever)
• Synchronize production rate with long term average demand
• Swim wear
– Chase (the demand) strategy (using capacity as lever)
• Synchronize production rate with demand
• Fast food restaurants
– Time flexibility strategy (using utilization as lever)
• High levels excess (machine and/or workforce) capacity
• Machine shops, army
– Tailored strategy
• Combination of the chase, level, and time flexibility strategies
Aggregate Planning
• Aggregate planning involves aggregate decisions rather
than stock-keeping unit (SKU)-level decisions for a
medium term planning horizon (2-18 months)
All-Terrain
Vehicle (ATV)

Engine
Transmission
Assembly

Model A Model B Model C Automatic Manual

Case Study Results
• In general, the chase strategy is used when
– Products are valuable
– Products are bulky or hard to store
– Products are perishable or carry an appreciable risk of
obsolescence
– High variety
• Accurate sales predictions are hard to obtain making stockpiling
hazardous
• Fashion items
• In general, the level strategy is used when
– Operators take a long time to become proficient at critical tasks
– Products with negligible probability of obsolescence
– Low variety
• Forecasts are quite good
Importance of Aggregate Planning

Without a sufficiently long-term view one

may make short-term decisions that hurt
the organization in the long-term
Importance of Aggregate Planning

• Aggregate planning at Henry Ford Hospital

involves matching available capacity, workers,
and supplies to a highly variable customer
demand pattern
Importance of Aggregate Planning

• Aggregate planning at Henry Ford Hospital

involves matching available capacity, workers,
and supplies to a highly variable customer
demand pattern
– 903 beds arranged into 30 nursing units
– Cost $5,000 of turning away a patient (simple cases)
– Cost of one idle 8-bed module is $35,000/month or
$420,000/year
– High degree of demand variability
• Demand for beds could change by as many as 16% in less
than two weeks
Importance of Aggregate Planning

Without a sufficiently long-term view one

may make short-term decisions that hurt
the organization in the long-term

• Shortly after Henry Ford Hospital reduced staff, it

determined the staff was needed
– New staff was recruited
– Both staff reduction and recruiting costs were incurred
Aggregate Planning

• Aggregate planning
– A general plan that determines ideal levels of capacity,
production, subcontracting, inventory, stockouts, and
even pricing over a specified time horizon (i.e. planning
horizon)
• Production rate (number of units to produce)
• Workforce (number of workers needed)
• Overtime (number of overtime hours)
• Machine capacity level (machine capacity needed)
• Subcontracting (subcontracted capacity)
• Backlog (total demand carried over to future periods)
• Inventory on hand (total inventory carried over to future periods)
Example: Aggregate planning at
RedTomatoTools
• RedTomatoTools
– A small manufacturer of gardening equipment

Shovels

Spades Generic tool, call it Shovel

Demand forecast

4,000

Forks 3,000

2,000

1,000

0
1 2 3 4 5 6
Inputs of an Aggregate Plan
• Demand forecast in each period
• Production costs
– labor costs, regular time ($/hr) and overtime ($/hr)
– subcontracting costs ($/hr or $/unit)
– cost of changing capacity: hiring or layoff ($/worker) and cost of
adding or reducing machine capacity ($/machine)
• Other costs
– Labor/machine hours required per unit
– Inventory holding cost ($/unit/period)
– Stockout or backlog cost ($/unit/period)
• Constraints
– Limits on overtime, layoffs, capital available, stockouts and
backlogs
Example: Red Tomato Tools

• Constraints
– Workforce, hiring, and layoff constraints
– Capacity constraints
– Inventory balance constraints
– Overtime limit constraints
– Inventory at end of Period 6 is at least 500
– Stockout at end of Period 6 equals 0
Example: Red Tomato Tools

Table 8-1
Month Period Demand Price
January 1 1,600 40
February 2 3,000 40
March 3 3,200 40
April 4 3,800 40
May 5 2,200 40
June 6 2,200 40

Aggregate plan decision variables

t Ht Lt Wt Ot It St Ct Pt
Month Period Hired Laid off WorkforceOvertime Inventory Stockout SubcontractProduction
December 0 0 0 80 0 1000 0 0
January 1 0 0 0 0 0 0 0 0
February 2 0 0 0 0 0 0 0 0
March 3 0 0 0 0 0 0 0 0
April 4 0 0 0 0 0 0 0 0
May 5 0 0 0 0 0 0 0 0
June 6 0 0 0 0 0 0 0 0
Average Flow Time

• Average flow time

– Average time one unit spends in inventory

Average inventory
Average flow time =
Throughput
Average Inventory

Average Inventory = (0.5(I0 + I1) +

0.5(I1 + I2) +
t It
0.5(I2 + I3) +
Month Period Inventory
Decemb 0 0
0.5(I3 + I4) +
January 1 1983 992 0.5(I4 + I5) +
February 2 1567 1775 0.5(I5 + I6))/6
March 3 950 1258
April 4 0 475
May 5 117 58
June 6 500 308
Average Inventory

Average Inventory = (0.5I0 + 0.5I1 +

0.5I1 + 0.5I2 +
t It
0.5I2 + 0.5I3 +
Month Period Inventory
Decemb 0 0
0.5I3 + 0.5I4 +
January 1 1983 0.5I4 + 0.5I5 +
February 2 1567 0.5I5 + 0.5I6)/6
March 3 950
April 4 0
May 5 117
June 6 500
Average Inventory

Average Inventory = (0.5I0 + 0.5I6 +

I1 + I2 + I3 + I4 + I5)/6
t It
= (0.5(I0 + I6) + I1 + I2 + I3 + I4 + I5)/6
Month Period Inventory
Decemb 0 0 0
January 1 1983 1983
February 2 1567 1567
March 3 950 950
April 4 0 0
May 5 117 117
June 6 500 250
Average Flow Time

• Little’s Law
Average inventory
Average flow time =
Throughput

1 T 1 1 1 T 1

Average Inventory   ( I t 1  I t )   ( I 0  I T )   I t 
T t 1 2 T 2 t 1 

1 T
Average Throughput   Dt
T t 1