Development of a decision support tool for Development of a decision - - PowerPoint PPT Presentation

• M. Grunow
• M. Grunow

Technical University of Denmark Technical University of Denmark Copenhagen Copenhagen ISORA 2008, Lijiang, China October 31, 2008

Development of a decision support tool for Development of a decision support tool for supply network planning: supply network planning:

H. H.-

• O. G
• O. Gü

ünther nther

Technical University of Berlin Technical University of Berlin Germany Germany

A case study from the chemical industry A case study from the chemical industry

CEEC 51

24.4

European Union 441 USA 393

14.4 4.1 12.8 7.3 6.5 3.5 10.4 5.8 4.8 17.3 3.3 6.7 5.5 12.2 4.6 7.4 19.0

Japan 188 Asia 183

Sources: UNSTAT Comtrade & Cefic-ITC Analysis

In billion €

World network of chemicals trade flows World network of chemicals trade flows

Latin America 71

• Large number of production sites

located in different countries

Supply network in the chemical industry Supply network in the chemical industry

Application background: intra-organizational supply network

• Production of a chemical

specialty

• Numerous industrial customers

across the globe

• Contracted annual quantities,

however, highly variable short- term replenishment quantities

• Introduction of a central supply

chain unit for coordinating

• Development of a customized

linear optimization model

• DEGUSSA AG, Düsseldorf, Germany

Company profile Company profile

• World largest producer of special chemicals
• Subsidiaries in all continents
• 46.000 employees / 300 plants worldwide
• Turnover in 2005: 11,800 billion €

Carbon black

Supply network in the chemical industry Supply network in the chemical industry SNP

• Lack of coordination results in

Excessive inventories Poor utilization of capacities Violation of delivery dates

• Supply Network Planning

Mid-term coordination of plant operations Integration of production and distribution activities

• Main planning tasks

Allocation of production volumes to plants Determination of the production mode for the main product and the generation of energy

Advanced Planning Systems (APS) Advanced Planning Systems (APS)

Strategic Network Design Supply Network Planning Demand Planning External Procurement Production Planning / Detailed Scheduling Transportation Planning / Vehicle Scheduling Order Fulfilment ATP / CTP

Mathematical methods

LP and MILP, heuristics

Decisions

Allocation of production quantities between plants Supply from the plants to DCs and customers

Based on generic model formulation

Generic model formulation

∑ ∑

∈ ∈

≤ ⋅

) ( ) ( i P p it pijt pi i J j

PC x a

∑ ∑ ∑ ∑

∈ ∈ ∈ ∈

⋅

) ( ) ( i P p I i pijt T t pij i J j

x c

∑

∈

≤ ⋅

) (i P p j pjt p

SC y α

∑ ∑ ∑ ∑

∈ ∈ ∈ ∈

≤ ⋅ + ⋅

) ( ) ( ) ( ) ( i P p j k i P p pjkt p j K pijt p j I i

HC z x α α

∑

∈

≤ ⋅

) (i P p ijt pijt p

TC x α

∑

∈

=

) (k J j pkt pjkt

b z

∑ ∑ ∑ ∑ ∑ ∑ ∑

∈ ∈ ∈ ∈ ∈ ∈ ∈

⋅ + ⋅

P p J j pjkt T t pjk ) j ( K k P p J j pjt T t p

z c y h

Assignment of attributes to pre-defined entities

Production capacity at plants Storage and handling capacity at DCs Transportation capacity per link Production costs per plants Storage and handling costs per DC Aggregate demand per DC

Advanced Planning Systems (APS) Advanced Planning Systems (APS)

Production process Carbon Black Production process Carbon Black

• Production volume: several 100,000 t per annum
• 100 product specifications
• Continuous production process

Feed A Feed B Feed C

Process

Main product Silo Gas Trans- formation

Energy

Silo Silo

Implementation of an APS Implementation of an APS Company negotiates annual volumes with key customers.

Forecasting the period distribution

• f annual demand

Forecasting the Forecasting the period distribution period distribution

• f annual demand
• f annual demand

Allocation of Allocation of customer demands customer demands to production sites to production sites

Customers request deliveries upon short notice.

Demand planning Demand planning

2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Month

t

Observed demand Forecast incl. seasonality

1997 1998 1999

• Adaptation of Winters forecasting technique
• Considerably increased forecast accuracy
• Forecast represents network-wide commitment

Forecasting the Forecasting the period distribution period distribution

• f annual demand
• f annual demand

Allocation of customer demands to production sites Allocation of Allocation of customer demands customer demands to production sites to production sites

Decisions in supply network planning Decisions in supply network planning

• Transportation volumes between production sites and customers
• Production volume in the production

sites and at each production train

• Generation of energy from side-products

Objectives in Supply Network Planning Objectives in Supply Network Planning

Transportation costs (Distance, carrier) Inventory costs

t , 1 Train , GER , p

PC

, 2 Train , GER , p

PC

t , 3 Train , GER , p

PC

t , 1 Train , RSA , p

PC

t , 2 Train , RSA , p

PC

Production costs (site, train) (negative) energy refund

Minimization of Minimization of Minimization of

Customer Demand

Limited substitutability of products

Delivery only from sites, which have a customer approval for the product A customer demand may only be covered by deliveries from a limited number of sites v p c t v p c A S s s t v p c s

CD A T

, , , , , } p) c, , s ( | { , , , ,

⋅ ≥ ∑

∈ ′ ∈ ′ ∈

p , c , s v , p , c t , v , p , c , s

B CD T ⋅ ≤

APS Implementation: Body of constraints APS Implementation: Body of constraints

Production

Distribution of the delivered

volumes at the production trains

APS Implementation: Body of constraints APS Implementation: Body of constraints

∑ ∑

∈ ∈ ′ ∈ ′ ∈ ∈ −

− + =

V v A p) , c (s, C c c t v p c s s TR s tr t p s tr s 1 t p s t p s

T PV SV SV

} | { , , , , ) ( ) ( , ), ( , , , , ,

Storage capacity

t p s t p s

MaxSV SV

, , , ,

≤

Production

Production capacity at equipment level

APS Implementation: Body of constraints APS Implementation: Body of constraints

t s e s e TR s tr P p p s tr s t p s tr s

AE O PV

), ( )) ( ( ) ( , ), ( , , ), ( ,

≤

∑

∈ ∈

APS Implementation: Body of constraints APS Implementation: Body of constraints

Production

Production capacity at equipment level

s T t V v P p A p) , c (s, C c c t v p c s

MinT T ≥

∑

∈ ∈ ∈ ∈ ′ ∈ ′ ∈ , , } | { , , , ,

APS Implementation: Body of constraints APS Implementation: Body of constraints

Production

Minimum sales from production sites

Energy

Transformation capacity of the sites

Transformation

t s s s TR s tr P p t p s tr s p s tr s

AT MaxCP PV CP

, ) ( ) ( , , ), ( , ), ( ,

⋅ ≤ ⋅

∑

∈ ∈

APS Implementation: Body of constraints APS Implementation: Body of constraints

Energy

Distribution of the aggregate energy of the various

energy types

Transformation Transformation

∑

∈ ∈

⋅ ⋅

) ( ) ( , , ), ( , ), ( , ), ( , s TR s tr P p t p s tr s p s tr s p s tr s

PV CP EV

∑

∈

⋅

ET et et s t et s

EET G 3600

, , ,

=

APS Implementation: Body of constraints APS Implementation: Body of constraints

Transformation Transformation

2 et s t 2 et s t 2 et s 2 et s t 2 et s

Y MaxE G Y MinE

, , , , , , , ,

⋅ ≤ ≤ ⋅

APS Implementation: Body of constraints APS Implementation: Body of constraints

Energy

Lower and upper limit on energy sales of the various

energy types

Graphical user interface User and scenario administration Solver (CPLEX) MILP models (OPL)

COM- Option SQL

Data

SQL

Data base OPL Studio

Optimization software architecture Optimization software architecture

Application of the optimization model Application of the optimization model

• Currently numerous managers are using

the tool for operative planning.

• Rolled out in Europe, US and Asia.
• Estimated financial benefit

per year from supply network planning exceeded project costs by far.

• Scenario mode is used extensively, e.g., for

capacity decisions, evaluation of approvals and

• f the profitability of energy transformation.
• Further benefits arise from

improved supply network design.

for for Thank Thank you you attention! attention! your your