Lecture 27:
Supply Chain Scheduling 1
© J. Christopher Beck 2008
1
Outline

The Beer Chain


Carlsberg Denmark Supply Chain
Management
Supply Chain Introduction



Strategic, Tactical, Operational
Planning vs. Scheduling
Hierarchical Decomposition
© J. Christopher Beck 2008
2
Readings

P Ch 8.6, 8.1, 8.2
© J. Christopher Beck 2008
3
Supply Chain Scheduling
© J. Christopher Beck 2008
4
Carlsberg


Sells many different brands
of beer
Sells many different “formats”


bottles, cans, kegs
6-pack, 12, 24
© J. Christopher Beck 2008
5
Carlsberg Supply Chain
Brewery 1
4 production
lines
Distribution
Centre
Brewery 2
2 production
lines
Warehouses
Stage 1
© J. Christopher Beck 2008
Stage 2
Stage 3
6
Stage 1 Scheduling

3 production steps on each line




brewing (and fermentation)
filtering
filling – bottling/packaging
All are resource constrained but filling is
usually the bottleneck

Filling operation has different costs and
processing times
© J. Christopher Beck 2008
7
Stage 1 Scheduling


All orders have fixed “lot size”
Products are divided into A,B,C
categories



A – high runners (a lot of demand)
C – specialty beers: more expensive, less
demand
Sequence dependent changeovers!
© J. Christopher Beck 2008
8
Stage 1 Transportation


Either to DC or direct to a warehouse
Different lot size constraints (truck
capacity)
© J. Christopher Beck 2008
9
Carlsberg Supply Chain
Brewery 1
4 production
lines
Distribution
Centre
Brewery 2
2 production
lines
Warehouses
Stage 1
© J. Christopher Beck 2008
Stage 2
Stage 3
10
Stage 2 & 3 Optimization



Placement of pallets at DC and
warehouses
Transportation to warehouses
Transportation to customers

vehicle routing
© J. Christopher Beck 2008
11
Scheduling Process

Medium term: 12 weeks


given demand and forecasts for products
3 MIP models solved sequentially

Costs: production, storage (at brewery, DC,
warehouse), transportation, tardiness, nondelivery penalty, and violation of safety
stock
Each MIP is composed of 5-10 subproblems
based
on
products
© J. Christopher Beck 2008

12
Safety Stock

One goal is customer service



Usually achieved by maintaining inventory
at DC and warehouses
Minimum inventory levels = safety stock
A lot of safely stock  good customer
service, but also high inventory costs!

(and skunky beer)
© J. Christopher Beck 2008
13
Short Term Scheduling

Based on medium term schedule, short
term scheduling plans the actual
production for one week



More detailed model of resources (i.e.,
sequence dependent setup costs)
Uses genetic algorithm or constraint
programming
Transportation scheduling
© J. Christopher Beck 2008
14
Overall Process

Medium term plan is re-done every day
using up-to-date information


takes 10 to 12 hours
Then short term scheduling is re-done
© J. Christopher Beck 2008
15
Comments


This is not a single model
Decompositions are crucial



medium term/short term
product-based
transportation scheduling decoupled from
production scheduling
© J. Christopher Beck 2008
16
Supply Chain
Optimization

Assume we are interested in minimizing
the cost of the entire supply chain


Individual participants will cooperate to
minimize overall cost
How many things are wrong with this
assumption?
© J. Christopher Beck 2008
17
Levels & Horizons
Level
Horizon
Strategic
1 – 5 years
Types of Decisions
Facility location, new products
Tactical
2 – 6 months Sourcing, distribution, orders
assigned to plants
Operational 7 to 21 days Production & transportation
scheduling details
© J. Christopher Beck 2008
18
Planning vs. Scheduling
Planning
Horizon
Scheduling
Multiple stages,
medium term
Information Aggregate
One stage/facility, short
term
Detailed
Objective
Time (e.g,. tardiness,
makespan)
Money
© J. Christopher Beck 2008
19
Hierarchical Decomposition


Planning solves higher level problems
based on aggregate data
The planning decisions are then used as
constraints (e.g., due dates) for the
scheduling


May be multiple independent scheduling
problems
Planning decouples scheduling problems!
© J. Christopher Beck 2008
20