|
Case study - BASELL Supply Chain Optimization for Polypropylene
Business
BASELL is the world's largest producer
of polypropylene, a leading supplier of polyethylene and advanced
polyolefin products, and a global leader in the development
and licensing of polypropylene and polyethylene processes
and catalysts.
Problem
The Polypropylene-Business currently has over 10 supply points
located in different European countries such as Benelux, Germany,
France, Italy, Spain, United Kingdom.
BASELL sells about 1500 different finished products belonging
to a number of different product families like homopolymer
products, impact and random copolymer products, metocene,
adstif, and clyrell products.
About 3000 different customers for these products are located
in Europe and Overseas. Customers are in the film, packaging,
car supplier, furniture, house ware, paint and other industries.
Every month, a 3 month sales forecast is provided by the BASELL
sales department. This forecast gives expected sales quantities
per customer, product, package, and month. Sales quantities
are qualified by customer categories that distinguish between
customers that must be served from sales opportunities where
optimization might or might not decide to serve depending
on profits that can be achieved.
The problem is to find a sales plan and a production plan
that takes into account:
| • |
Given minimum
and maximum sales quantities |
| • |
Production
and stock capacities |
| • |
Opening
stock quantities |
| • |
Production
constraints such as minimum lot sizes, fixed lot sizes
("batches"), raw material availability and more |
| • |
All relevant
costs such as raw material prices, production costs, inventory
costs, transport prices |
| • |
Sales prices
per customer, product, and package |
The objective function maximizes the
profit: sales income minus costs.
Implementation
The user interface is implemented in MS Access and has interfaces
on one hand to the company's SAP systems, on the other hand
to the Xpress-MP optimizer.
A typical planning session comprises the following steps:
| • |
Load current
data from SAP: material and customer master data, production
data (recipes, raw material costs), sales data (forecast
and sales prices) |
| • |
Verify correctness
and completeness of the data: Do all expected sales have
a potential source? Are the raw materials needed during
different production steps available? Are there any contradictions
in the constraints? and others. |
| • |
Run the
optimization and look for "flaws" in the results.
Such "flaws" could be: stock quantities becoming
higher or lower than desirable; unbalanced plant utilization
rates; erroneous input data leading to "strange"
results; and others. |
| • |
Define additional
constraints in order to avoid "flaws" described
above. Repeat optimization until a satisfactory result
is found. |
| • |
The final
result is documented in reports and sent back to SAP as
basis for (1) production scheduling executed at the different
production plants and (2) reservation of sales quantities
for the specified customers. |
The system is built in a way that different
versions of the current data can easily be generated and administered,
allowing to create "what-if" analyses and to go
back to historical data.
Details
| • |
In every
planning session, about 800 different products, 1,500
different combinations of product and production plant,
10,000 different combinations of customer, product, package
and month are involved. |
| • |
These numbers
lead to over 200,000 variables, 380,000 non-zero elements,
400 integer variables and 900 semi-continuous variables. |
| • |
Solution
time for the linear solution is about 1 minute; for a
"good" integer solution, that takes into account
semi-continuous and integer variables like minimum lot
sizes and fixed lot sizes and that is guaranteed to be
within 5 % of the best possible solution, solution time
is about 15 minutes. |
Summary
The optimization system described above was first implemented in
1996, where it had to deal with 4 production plants and about 100
different final products only. Through various acquisitions and
mergers of the company, the system grew to the size it has today.
The BASELL planning department says: "It is of major importance
for us to have an instrument that gives us the ability to integrate
the complex planning and logistics process. Our optimization systems
for Polypropylene Standard Products have become a key element in
production and logistics planning. During the last big merger in
2001, optimization had a most prominent role: merge two polypropylene
suppliers by first merging their planning activities!"
|
|
